I understand that utterances on social media, often in response to comments by others or made in anger, are often quickly regretted, and I believe (though I am not certain) that Mr. Khosla did not quite mean what he said here, confusing profitability with cash flows, and arguing that every business should put scaling ahead of profitability. That said, his view that scaling should be given priority over profitability is more the norm, than the exception, among many venture capitalists, and while it probably always has been the case, I believe the tilt towards scaling has become pronounced in the last two decades. In this post, I want to zero in on the scaling and profitability trade off, how the emphasis on the former over the latter plays out at start-ups and very young companies, and why we live with the consequences, whether they want to or not.

Scaling versus Business Building

To put the choices you will face on scaling up versus business building into perspective, let’s assume that you are a founder, and that your start-up has a tested product and that you believe there is a market for that product. You can stay with what you have built and build a business to take advantage of the immediate market, focusing on financial health and profitability. The fact that you will stay small, and perhaps unrecognized in markets other than your own, is a minus, but there are pluses. You will have little need for external capital, and you will own much or all of the business, facing little pressure from outside to change the way you do things. Alternatively, you can take a more ambitious route, where you seek out a bigger market, augmenting existing or adding new products, and while that path will deliver larger revenues, you may have to work harder to get it to deliver profits and cash flows, and perhaps have to give up more of your ownership and control of that business.

The Scaling Choice

Before starting on the determinants of scaling, it make ssense to begin with the metric being scaled. For most businesses, it is revenues that is the chosen metric, with scale capturing how big revenues can become over time. With some earlier-stage businesses, many of which are pre-revenue, the metric can become a variable that these businesses hope to convert to revenues; with tech intermediaries and social media companies, it can be users or subscribers.

Focusing on scale, though, there are factors that come into play that allow scaling to have a higher likelihood of success in some businesses than others:

1. Market size: It is easier to scale up a company, if it is small player in a big market, than if if the market is small, and scaling up will quickly give you a dominant market share. That said, the way you describe your business, and then run it, can play a role in how big a market you will have for your products. In my posts on valuing Uber, for instance, I noted that describing it as a logistics company (car service, moving, delivery) rather than just a car service company could triple its potential market.

2. Market growth: It is also easier to scale up a company if the overall market that it is targeting is also growing, since growth does not require going after competitors’ customers. A smartphone company (Apple or Samsung, for instance) in 2010 had a growing market to work with, as customers switched from flip phones and smartphones made inroads into large emerging markets. In 2026, that advantage had largely dissipated, as the smartphone market has matured.

3. Industry Structure: There is a natural structure to industries, driven by economics and business type, with some industries splintered across many players, and some concentrated in a few big players or even in a winner-take-all. You can scale up more in the latter, but you will have to confront the odds favoring you being one of the winners in the industry.

4. Capital intensity: It is easier to scale up a business that does not require large capital investments to be able to generate more in revenues. Using Uber as an example again, scaling up was made easier in the early years, since it did not own the cars or hire the drivers that comprised its car service, and growth came quickly and with little added investment.

5. Customer inertia: Businesses can grow faster and get bigger if there is less inertia among customers and more willingness to try out new products or services. At the risk of generalizing, this may explain why scaling up can happen more quickly in younger industries (like technology) than in older ones (health care, education).

6. Key person(s): There are some businesses that are built around the specific skill sets of a person (usually a founder or business owner) and these skill sets are not easily transferred or taught to others. A master craftsperson, say a furniture-maker, will have a more difficult time scaling up that business, because without being able to pass his skills on to his or her apprentices (which can take time and require intense oversight), he or she is constrained in how much new business he can take on. If that craftsperson has a recognizable name, it is possible that you could build a scalable franchise model, as has been tried by some master chefs (Wolfgang Puck, Gordon Ramsey etc.)

The graph below captures the scaling choices that companies make as a function of these factors:

As you can see, some businesses can scale up quickly, some take more time to scale up and some never scale up, and the businesses that scale up quickly often scale down just as fast. Thus, the decision of whether to scale and how quickly to do so is as much driven by the nature of the business (capital intensity, industry structure, competition) and the characteristics of the market that it is targeting (size and growth, customer inertia).

Business Building

While having access to a big, growing market can allow you to scale up more quickly, your capacity to generate profits and build a business will ultimately come from other forces:

1. Unit economics: Unit economics measures the profitability of the marginal unit sold by a business, and is thus determined by the price charged for that unit and what it costs the business to produce that unit. Businesses like software, where the marginal unit costs very little to produce and can still be priced highly, have superior unit economics and will find it easier to convert growing revenues into profits, since much of the increase in revenue will flow into profits. Conversely, businesses like electric cars, where each additional car sold costs money to make, will struggle to convert scaled up revenues to profits.

2. Economies of scale: Businesses with large fixed costs, whether they be associated with maintaining platforms and infrastructure, or sales and marketing, face obstacles to profitability. While growing can provide scaling benefits, that works only if the fixed costs don’t grow with revenues and if they are not so onerous, that you still have losses after scaling up.

3. Competition: & Competitive Edges (moats): Large and growing markets provide businesses with opportunities to grow, but for that growth to translate into sustainable profits, these businesses will need pricing power and that power comes from barriers to entry that keeps new entrants out and gives existing players advantages.

It is true that the operating choices that businesses make play out on both the scaling and profit dimensions, sometimes pitting them against each other. A decision to lower product prices may increase revenues at the expense of unit economic profits, and a decision to spend more on advertising and promotion may expand markets, but the higher marketing costs will impose a drag on profitability.

One way to illustrate the combination of forces that go into business building is to to go back to basics, and to look at what lies under each one:

As you can see, scaling up is not a mantra that automatically translates in profitability, and the pathway to profits will be determined by variables that are often out of the control of a business.

Scale & Profitability Mixes

With the multitude of factors determining both scaling potential and business model viability, it should come as no surprise that the outcomes that we observe can range the spectrum, starting with extraordinary companies that scale up quickly, while delivering huge profits, to companies that never scale up, either by choice or because they could not, and some of which never make money.

1. Lightning in a Bottle: Are scaling and profitability mutually exclusive? Put differently, can a company scale up, while delivering profits and perhaps positive cash flows as it grows? The answer is yes, but it does require a fairly unusual combination of circumstances - a big and growing market, being an early entrant into the market with few competitors, low capital intensity and excellent unit economics. There are a few companies that meet these conditions, and we will call them “Lightning in a Bottle” firms, partly because they are rare, and partly because success can come from being at the right place at the right time. Google and Facebook, in their early years, were good examples, with revenues growing exponentially and profitability in place.

2. Field of Dreams (Shoeless Joe Jackson version): As a baseball fan, I have always had a soft spot for the movie, Field of Dreams, where a farmer (Kevin Costner) builds a baseball field in the cornfields, and when asked why, responds with “if you build it, they will come”. There are companies that seem to be built around this motto, where scaling up comes first, often accompanied by large losses, but with the promise that “if they build (revenues), they (the profits) will come. During Amazon’s first decade and a half of existence, I described their business model as a Field of Dreams model, and gave credit for Jeff Bezos for being steadfast in not only telling this story, but also acting consistently with it, and carrying investors along. (If you are wondering what Shoeless Joe is doing in this story, I am afraid you have to watch the movie all the way to the end.)

3. Field of Nightmares: Amazon was not the first successful Field of Dreams company, but as one of its highest profile winners, it gave rise to a legion of young companies, all labeling themselves the “next Amazon”. Needless to say, Amazon’s success came from being a disruptor of a huge business (retail), which had atrophied and weakened over time, and many of the Amazon wannabes that tried to imitate it managed to do so on the growth dimension, with immense amounts of capital invested in scaling up, but never turned the corner on profitability, partly because they had neither the unit economics nor the economies of scale to pull it off.

4. Niche Star: Scaling is not always the optimal choice, and there are some companies that recognize this reality early, choosing to stay small and focusing on a portion of the market where they have decided advantages. To that extent that they can convert those advantages into premium pricing and niche market dominance, they can have values that are disproportionately large relative to their operating metrics, i.e., trade at high multiples of revenues and earnings. Ferrari, for instance, sells only a few thousand cars every year, but with an operating profit margin in excess of 20%, it trades at a market capitalization comparable to that of auto companies that sell hundreds of thousands of cars each year.

5. Big and Broken: It is no secret that there are some businesses that start with business models with a fatal flaw, i.e,, a broken business model, and rather than being shut down, they are fed increasing amounts of capital and allowed to scale up. A real-estate based business that leases properties long term, and then sub-leases them short term, has a duration mismatch born in hell, and expanding it geographically and allowing it to lease hundreds of properties, as WeWork did, just makes it a really big, bad business. If you are puzzled as to why investors would supply capital to these businesses, you may want to read on.

6. Small winner & Small losers: If you look at all businesses, private and public, most remain small, some due to business and industry structure and some because of owner constraints on capital and control. These small businesses, though, over time, bifurcate into good small businesses, earning more than their cost of capital and delivering value, and bad ones, earning less than the cost of capital, but still worth more as going concerns, than liquidated.

7. Cut your losses: Finally, there are businesses that start up with dreams aplenty, and over time discover that they can neither scale up, nor make money. In the absence of capital infusions, these businesses fail early, but if capital providers keep funneling resources into these companies, they still fail, but do so later and with a much higher price tag.

In the matrix below, with scaling on one axis and profitability on the other, I plot all eight of my scale/profit combinations:

Any investor or founder who blindly follows the pathway of scaling first and profiting later for every business is using a cookbook approach to business building, and runs the risk of making small failures into big ones.

The Tradeoff between Scaling and Profitability: Determinants

As you review the factors that govern the trade off between scaling and profitability, it is clear that the right choice (on how much to scale) will depend on the firm, and that not every small firm is destined to become or be more valuable as a larger firm, and that not all large firms have the same profitability characteristics, once scaled up. That said, is it possible for firms to adopt scaling pathways that look, at least from a business standpoint, to be suboptimal? Of course! There are small firms that have viable pathways to scaling up that choose to stay small, and at the same time, there are small firms that are designed to be small, niche businesses embark on scaling that is value destructive, and the reasons are a mix of human frailties on the part of founders, system constraints (from governments and regulators), access to capital (too little or too much) and exit options (sell, liquidate or go public).

1. Founder Characteristics

The founder or founders of a business not only play a key role in guiding the business through its early days, when most start-ups fail, but they also make key choices that can determine in its end game. In making these choices, they may be guided by the fundamentals we outlined in the last section, that affect scalability, but they are also a function of their personal make-up, on at least a couple of dimensions:

• Control versus Ambition: There is a natural tension between wanting to control the levers of decision-making in a business and scaling that business, since the latter almost always requires raising capital from providers who will either constrain your choices (if borrowed money is used) or demand a share of ownership rights (if equity). With the latter, founders will find their control diluted over time, and with enough scaling up, it is possible that founders end up with less than controlling stakes. For some founders, that fear of dilution and losing power over their business creations runs deep enough to stop them from embarking on growth plans, even though these plans make economic and financial sense.The flip side of control is ambition, and for some founders, the desire to build big businesses that are not restricted geographically or in product offerings can drive the decision to scale up, even though the fundamentals may not support that expansion. This works only if they can convince investors that their ambitions In fact, this tension between a founder’s need to be in control and that same founder’s desire to build big plays out in what Noam Wasserman called the Founder’s Dilemma, where to make a business bigger, its founder has to step down or at least compromise on control.

• Longevity versus Scale: There is an argument to be made that if your intent as a founder is to build a business that is long-lived, your odds of success improve if you keep your business smaller and more focused on what it does well. While there are many exceptions to this generalized rule, it is worth noting that some of the longest lived firms in the world are family owned small businesses, that serve a niche market, and are passed down generation to generation in the same family. It is also true that firms that see a sudden surge in revenues, usually as the result of an external factors or happenstance, often live to regret their good fortune, as they scale up overnight. In the aftermath of the Covid shutdown, for instance, firms like Moderna and Peloton boomed, but they also overreached, and did long-term damage to their business models.

In summary, the choice between scaling and profitability will play out differently across businesses, depending upon what founders value most, thought it is healthy for an economy to a have a mix of founders, since it creates a mix of businesses.

II. Access to capital

It is true that businesses need access to capital, to varying degrees, to scale up, and the easier it is to raise that capital, the easier it is to make a business bigger. Capital can come from different sources, ranging from family wealth to venture capital to public equity, with each one carrying its pluses and minuses.

• Family (or friend) wealth: Every business, through human history, having lived through its early days (when failure risk is high and its products and services are still untested) has faced a choice of whether to stay small, serving a market that it knows and understands, or whether to get bigger, going after a bigger market. For much of that history, though, with businesses funded with family funds and access to capital was limited, most businesses chose the first path and remained small businesses, focusing on building business models that delivered profits, with wide differences in success rates. For a few, owned by wealthier families, access to a much larger pool of capital (from family savings and bankers willing to lend to these families) created family groups that dominated economies, and continue to do so in some parts of the world.

• Venture capital: The growth of public equity markets in the late 1800s and much of the last century did little to change the family control dynamic, since investors in those markets were primarily interested in funding larger companies with established business models. Recognizing this gap between capital need and capital access at younger businesses, and the opportunities that the gap presented, allowed for the rise of venture capital in the 1950s, primarily in the United States. These venture capitalists provided seed capital for start-ups, using winners to cover their failures, and got the bulk of their winnings when they exited these investments, either by going public or selling to another entity. Over the last few decades, venture capital has grown, and in the last 12 years, that growth has not let up:

  

Source: NCVA 2026 Yearbook

In this century, venture capital has also become more global, growing in Asia and Europe, but it is still true that it is easier for a small business to raise capital to scale up in the United States than it is in much of the rest of the world.

• Public equity: There are some growth businesses that bypass venture capital and go after public equity, a much bigger pool of capital and one that may give founders better terms. In some cases, this access to capital might be enabled by going public, even with unformed business models and little to show in terms of existing operations (revenues or earnings), but in most others, it takes the form of capital invested by larger, more mature public companies in return for a share of ownership. These investments may be labeled as strategic, but the motives for making these investments vary across companies. Some invest to get access to a promising technology or product. some to pre-empt competitors and some for the same reason that venture capitalists do.

The bottom line is that businesses that seek out capital, whether from family, venture capital or public equity, have to accept that the capital providers will demand and usually get a say in business decisions, and the more capital you seek, the more sway they will have.

III. Investor Preferences

Businesses get their cues on whether to scale up or build business models from the investors who fund them, and much as founders want to map their own path, investor preferences matter, as do their end games. Put simply, a family that invests in a business with no plans for exit will choose a very different path for that business than a VC that invests in the same business with the intent of exiting that investment by selling it to another investor or company, or taking it public.

Venture capitalists are often viewed as the sherpas who guided young businesses to success, both operationally and in markets, the mythology about venture capitalists and what they do has also built up. Since that mythology extends to almost every aspect of venture capitalist activity, perhaps the best way to dispel myths and bring in reality checks is to look at what venture capitalists are “assumed” to do in each phase, and contrast it with what they actually do:

If you are reading this as a critique of venture capitalists, you are misreading it. My intent is not to paint a picture of venture capitalists as lazy and greedy, but to bring home the reality that given how venture capitalists invest, act and are judged, it is unrealistic to expect them to do the heavy lifting of building businesses for the long term and to even make business sense, when they talk about companies.

There are two parts of the venture capital rulebook that you should focus on, to understand why many VCs prioritize scale over profitability. The first is that they price companies, rather than value them, and in a post from a few years ago, I made the argument in more depth. VC pricing based on what other venture capitalists are paying for similar businesses, often scaled to simplistic metrics, users and subscribers for pre-revenue companies and forward revenues or earnings in what passes for VC valuation:

The second is that VC success is measured based on price at entry and price at exit on an investment, rather than the quality of the business built, and using that metric, the median venture capitalist has not been much better at harvesting alpha than the median mutual fund manager or PE investor:

Cambridge Associates

There are, of course, standouts in each of these categories, fund managers who have delivered well above the market, but in mutual funds and to an increasing extent, hedge funds, that success is fleeting. There are two aspects on delivering returns where venture capital stands out, relative to other active investing classes.

• The first is that failure, always a concern in investing, is much more a part and parcel of investing in venture capital than in other investing grouping. Put simply, not only are there more VC funds that go out of existence every year, but even the most successful VC funds lose on many or even most of the investments that they make, especially in angel financing deals.

• The second is that venture capital investing, when it works, can generate outsized returns on winners that (hopefully) cover the cost of failures.

You can see both of these at play in the graph below, which looks at returns that VCs book when they exit investments:

CF Private Equity, from Pitchbook data

As you can see, across all the time periods, it is the top 10% of VC investments that deliver the bulk of returns to VC investors, and over time, that concentration has increased: in the 2023-2026 period, 80% of all returns to VC investors came from their top 1% of investments. The combination of these two forces (losses on most investments and outsized winners), i.e., the power law in venture capital, has two consequences. The first is that only about a quarter of venture capitalists in each year deliver above-average returns, making the average VC returns in the table above more palatable. The second is that success in venture capital, unlike in other areas of active investing (including mutual funds, hedge funds and even private equity), has been more enduring. The power law characteristic also feeds into VC incentives, leading venture capitalists to direct their capital more into chasing the biggest winners than in building businesses. In fact, the more top-heavy VC returns become, i.e., dependent on big payoffs, the more pressure venture capitalists feel (and pass on to their portfolio companies) to find the next big winner, pushing the ecosystem dangerously close to gambling.

A Changing Game

With the discussion of the scale versus profitability at the business level leading in, and the assessment of the incentives of capital providers following, I think that we are well positioned to examine how changes in public and private markets have increased business incentives to scale, as opposed to building business models. There are two developments, in particular, that have taken the tilt towards scaling in venture capital and made it even more pronounced - the entry of public equity into the funding of private businesses and the fading of reversal, as an antidote to momentum, in public markets.

The Gray Market Effect

For much of the last half of the last century, after venture capital established a presence in the United States, it remained the only or primary source of capital for young firms. That has changed especially int the last decade, as public equity investors have increased their investments in young, private businesses, supplementing venture capital in some and even displacing it in others. An early measure of this trend is captured in the charts below:

Kwon, Lowry and Yiming (2020)

While this graph looks at only the number of mutual funds investing in private businesses, and stops in 2016, there was a corresponding surge in capital invested by mutual funds in young, growth companies, with T.Rowe Price and Fidelity investing billions in high profile tech companies like Uber. They were joined by sovereign funds, who invested heavily in these companies either directly or indirectly, through stakes in entities like Softbank’s Vision fund.

We can debate the reasons for why we saw this surge, with fear over missing out (FOMO) and wanting to partake in tech playing roles, but whatever the reasons, capital access surged for young companies, especially in tech, during the period. In effect, rather than two mostly separated markets - one for young, smaller, private business dominated by VCS and one for larger companies more advanced in the life cycle, where public equity suppled the funds, a gray market was created where VC and public equity fund access allowed private businesses to stay private for longer.

Public Markets: Momentum, Fundamentals and Reversals

Public equity markets have always been momentum-driven, allowing traders who ride that momentum to prosperity, before bringing them down when the momentum shifts. At the same time, fundamentals act as an anchor, operating as a counter to momentum, leading to reversals and allowing investors to hold their own over time. While the congruence is not always perfect, scaling feeds into momentum and profitability is the most critical fundamental, and in markets with balance, when one gets out of sync, the other restores harmony.

Over the history of stock markets, value investors have often claimed dominance, and pointed to the returns you could have earned by buying companies that look cheap on a value basis (low price earnings or low price to book) and waiting for price reversals. Traders push back by noting that over the same history, momentum has had a decisive effect on returns, especially over shorter time intervals. While the momentum effect shows up across the decades, there is evidence that the reversal effect has weakened over time, leaving investors who bet on mean reversion and a return to fundamentals in the lurch:

Source: Ken French Data

The reasons given for this shift vary, and are often reflective of the biases of the investors giving the reasons.

1. The Fed did it: For those who view central banks as all-powerful, and believe that the low interest rates of the last decade were their doing, those low rates have also become the proximate reason for market pricing behavior and reckless risk taking. Their argument is that interest rates that are close to zero induce investors to shift from bonds to stocks, and within stocks, to move from low growth, high earnings stocks to high-growth companies with little or negative earnings.

2. The rise of passive investing: In the battle between active investing and passive investing, with ETFs supplementing index funds, the latter has had a decisive edge in terms of returns over the last two decades, and its share of the market now stands are well above 50%. There are some who argue that the flow of funds to passive investing vehicles has contributed to the increased power of momentum, since more new funds flow to the largest market cap companies than to the smaller ones. In addition, it is argued as the number of active investing declines, there are fewer investors looking at business models and profitability, reducing the pull of fundamentals on price.

3. Public market composition: It is noteworthy that the reversal effect started weakening in the 1990s, a decade when young dot.com companies with unformed business models flooded the market, bypassing the more traditional route of using venture capital to grow. With these companies, where value is almost entirely driven by potential and not by operating metrics today, the catalysts needed for reversal may take longer to manifest.

4. Information sources and access: It is undeniable that investors and traders get information from a wider ranges of sources now than two or three decades ago, with social media and online sources supplying information that used to come from newspapers and financial news channels. In additional to being less curated and controlled, that information is also instantaneously accessible to the public, and price reactions tend to follow.

While I take issue with parts of each of these arguments, there is some truth to all of them, and they have contributed to making pushing back against momentum a more hazardous exercise for investors.

The Consequences

With larger amounts of capital being deployed by VCs at young, growth companies, substantial capital infusions from public equity funds into private capital markets, and public equity markets that are more used to and receptive to young company listings, it should not be surprising that it is changing how private companies behave. In the graph below, I look at the characteristics of companies going public in the United States, using the data that is generously made available by Jay Ritter:

Source: Jay Ritter’s IPO data

There are three clear changes over time that are visible in this graph:

1. Private businesses are waiting longer before going public: As you can see, the average age of a company going public has risen over time, with the median age rising about 11 years in the last 15 years.

2. Private businesses are scaling up (revenues) more, while waiting: While private businesses wait longer to go public, they are spending that time scaling up more than they used to. The inflation-adjusted revenues at the median IPO have tripled or even quadrupled, relative to IPOs in the 1980s.

3. Private businesses are deferring building business models & profitability: The most striking feature of the data, to me, is that while private businesses are waiting longer and scaling up more before going public, they also seem to be deferring business building for much longer as well. While it was routine for companies going public in the 1980s to be profitable (>80% were), less that a quarter of the companies that have gone public in the last decade have been profitable.

While companies that are going public are bigger (in revenue terms) and less likely to be profitable, markets are attaching large market capitalizations to these newly minted companies, as you can see in this graph which zeros in on tech IPOs:

Source: Jay Ritter’s IPO data

You will also notice that companies going public are issuing smaller portions of their shares to the public, at least in the initial offering, suggesting that the need for capital that drove companies to go public has become less pressing over time, perhaps because of more capital access as private businesses. While the median market cap of a company going public in the last six years has exceeded a billion, the largest IPOs command market capitalizations that would have been unimaginable a few decades ago. From Facebook, with a pricing of $104 billion, in 2012 to SpaceX, going public in June 2026 at $1.8 trillion, the trend lines are pointing upwards, especially if Anthropic and OpenAI deliver on their trillion-dollar plus pricing promise.

Implications

By itself, the trend towards private companies scaling up more, while public, and going public at eye-popping market capitalizations may be understandable and explainable, but there are implications that we need to consider both from an investing and governance standpoint.

1. Corporate governance: One of the reasons that private companies often delay going public is because governance requirements, from board composition to top management compensation, are more stringent at public than private businesses. While Sarbanes-Oxley, which wrote into law many of the current governance rules for public companies, is often toothless and ineffective, it still forces disclosures about governance (on conflicts of interest and board member relationships) at public companies. In addition, public market investors can pressure public companies to change governance practices or top management, if companies underperform in the market place. One of the perils of letting companies scale up more before these governance questions get raised is that the top management in these companies may have few checks on their actions. It is true that venture capitalists could operate as a disciplinary mechanism, but in an age of founder worship and where VCs can be divided and conquered, you can have companies with market pricing of a billion, hundreds of billions or even trillions run by people who are ill-suited for the task.

2. Delayed business model building: If the first imperative for a private business is to scale up, because scaling pushed up pricing both in private and public markets, the challenge of business building will get deferred to a later stage. The problem with scaling up first, and building a business model later, is that it may be too late, since the choices made to allow for scaling up may impede the pathway to profitability. Again, if your response is that VCs will work on fixing this problem, they have little incentive to do so, since they benefit from scaling up and exiting these businesses, before the business problems become too big to ignore.

3. Scaling stories: If you believe, as I do, that valuation is a bridge between stories and numbers, and that the balance between the two shifts over the life cycle, with stories dominating early in the life cycle and the numbers taking center stage in the later stages, it is understandable that VCs and founders, when marketing their companies are primarily story tellers. I don’t have a problem with that, but as I noted in my last post on AI as a business, the stories that are being told for these companies are often incomplete, and almost entirely focused on the scaling question. Thus, in the Anthropic sales pitch it is the growth in the annualized revenue run rate (ARR) and the size of the AI market (huge, but with no specifics) that comprises the bulk of the story, with little or no mention of business models or profitability.

4. Disruption without replacement: Disruption has been a key component of the stories that underlie many of the largest companies that have gone public in this century. Accepting the premise that a healthy economy needs a shaking up of the status quo, and that disruption can lead to economic growth and better practices, it is still legitimate to look at disruption’s debris. One of the perils of supplying capital in almost endless quantities to private businesses that aim to disrupt, without challenging them on business models, is that you may succeed at disrupting the status quo (driving existing players out of business) but your disruptor may not be able to build a business that can be self-sustaining in the long term.

Conclusion

I am sure that you are already aware of the core message of this post, which is that notwithstanding the current emphasis on scaling up businesses, not all businesses are meant to scale up, and that scaling up comes with challenges that founders may be ill-equipped to meet. That said, ambitious founders will feel the urge to make their businesses bigger, and if they raise capital (from venture capitalists) to make this happen, the incentives to scale up will increase, even if it makes little or no business sense to do so, with all parties hoping to exit by selling to others (public or private) who will price based on scale. While this has always been the case, changes in private and public capital markets have tilted the scale even further in favor of scaling, and it is possible that companies, both public and private, with sky-high pricing have been built on bad business models that are irredeemable.

YouTube Video

Blog posts on Venture Capital and Scaling

1. Blood in the Shark Tank: Pre-money, Post-money and Play-money Valuations (February 2015)

2. Billion-dollar Tech Babies: A Blessing of Unicorns or a Parcel of Hogs (June 2015)

3. Venture Capital: It is a pricing, not a value game! (October 2016)

4. Risk Capital in Markets: A Temporary Retreat or a Long-term Pullback (July 2022)

1. Original post: August 20, 2026

2. First update: August 21, 2026 - Updated AI TAM table to make explicit implications of AI market size for other businesses, employees and the economy.

3. Second update: August 21, 2026 - Updated to incorporate impact of new businesses that may be born out of AI

In a world where AI enters almost every conversation, it takes effort to remember that its breakout moment was less than four years ago, when, on November 30, 2022, ChatGPT was unveiled to the public. I know! I know! Artificial intelligence has been around a lot longer, with a history tracing back to the birth of the computer age. I am old enough to remember IBM’s Deep Blue, a machine powerful enough to evaluate two hundred million chess positions per second, and beat the greatest chess players of its time. On the cultural front, we have seen variants of stories, where machines break free from human control and take over the world, in novels and movies, with Hal (the computer) in 2001: A Space Odyssey retorting “I’m afraid I can’t do that” to Dave, his human controller, remaining one of my favorite movie lines of all time.

Notwithstanding its longer history, the effect on AI has been explosive in the last four years, manifesting in multiple developments. The most successful company during this period, in terms of increasing market capitalization, has been Nvidia, the chip maker for the AI revolution. After spending a decade talking about FANGAM, the big tech companies that had become part of our daily lives while carrying equity markets forward over the last decade, it was the Mag Seven that became the stand-in for market dominance, with Tesla and Nvidia replacing Netflix in the mix. The Mag Seven, almost all of which have a stake in AI, have accounted for 45% of the increase in market cap across all US stocks between 2022 and 2025, and have an aggregate market cap on August 16, 2026, of $23.7 trillion. It is not just markets that are besotted with AI, since the massive investments in AI architecture, from data centers to large language models (LLMs) have carried the US economy; it is estimated that these investments accounted for about 1% of the 2.5% in real GDP growth in 2024 and 2025. Almost every conversation of businesses now has an AI component, which if not restrained, can hijack the discussion.

AI’s effects were not restricted to business and markets, as people were exposed to its reach in their personal and work lives, with reactions ranging from awe, at its power to do tasks that used to require skilled human labor, not just effectively, but in a fraction of the time, to dread, at the possibility of being made obsolete by an AI agent. As the arc of the AI story has unfolded over the last four years, it seems to me that is has also transitioned in the public consciousness from a mostly positive phenomenon early on to acquiring a negative tinge, perhaps because of concerns that the genie is out of the bottle, and is not benign, and partly because some of its leading spokespeople are so unlikeable. Not surprisingly, speakers at graduation ceremonies in US colleges in 2026, were booed by students, when their speeches centered around AI.

The AI Debate: Off the tracks?

As the AI story has unfolded, there have been reams written about it, for and against, and almost as much said about it on television and podcasts. In spite of being so much in the news, the debate about AI, in my view, has gone off the track with advocates and skeptics often talking past each other, with advocates focusing on its “huge” potential market, and skeptics zeroing in on massive upfront investments as “too large”, with each side claiming the high ground and labeling the other side as cultists (AI advocates) or Luddites (AI skeptics). There is a great deal of cherry picking of the data on both sides, with the optimists focusing on usage statistics (level and growth) and the pessimists on capital spending and current profitability (or lack of it).

This post is not about proving one side right and the other wrong, but about closing the loop and making it a discussion of AI as a business, recognizing that it is ground breaking, while also acknowledging that it has to be judged like every other business in history, not on potential usage, but on the prosaic details of converting potential to products and revenues, being able to deliver these products at a cost that generates profits and building moats to keep new entrants and competitors out. In short, the AI optimists may be right about AI usage exploding in the future, but big markets don’t always become big businesses, and the skeptics have to concede that spending a lot on capital expenditures raises the ante for businesses, but don’t necessarily doom them to value destruction.

I will be writing this as an AI novice, a very light user of ChatGPT (I still have only the free version) and acquainted with Claude only in passing (though my content has found its way into some of its bots). If you are an AI expert and feel that I am missing or wrong about a technical component, forgive my ignorance, and feel free to educate me, and if you work at an AI business and feel that I am in error on a business detail, the same offer stands. To be honest, I am writing this post for an audience of one (me), with the purpose of clarifying for myself how to make sense of this space, and if it does help you make sense of this disruption, it is a side benefit.

The Cycle of Revolutionary Change

Through human existence, revolutionary change has been a constant, and even when that change has been an advance for humanity, it has always come with pain for those that the change renders obsolete and unanticipated side costs. At the risk of overreach, I will argue that every major disruptive change has gone through four phases: a period of hope and hype, where the change is viewed as big, but it is unclear how and in what form it will be delivered, a period of build-up, where a subset of people (with more belief in the change and more willingness to take risks) start investing and building products to make the change happen, a period of business building, where the change is monetized and businesses form, and a recalibration, where the change works its way through the economy and society, in both good ways (increased productivity and welfare, new businesses) and bad ways (displacement and damage).

1. Hope and Hype

In this phase, the true believers and visionaries that see change coming start the ball rolling, but to succeed, they need to sell it to the broader public. Since the story of change, at least at this stage, has nothing tangible at its core (no products or services, let along revenues or profits), it is inevitable that there will be false starts mixed in, as well a dose of scams pushed by charlatans and pretenders. It is also par for the course that there will be many who will dismiss change talk as fairy tales, without even listening to the arguments, either out of cynicism or because they do not understand what is being sold. For change to take root, the visionaries selling the story need to be persuasive enough to get people to buy into their vision, both to get foot soldiers who will work to make change happen and investors to supply them with capital.

2. The Investing Build-up

Once the belief that change is possible gets a foothold, there will be a subset of players, with start-ups or in existing businesses, that will invest and build products and services that they believe will be sought after, if change comes. As pioneers in this space, with trial and error and experimentation characterizing these attempts, but even failures will lead to learning, albeit with costs.

If the change is perceived as revolutionary, with a big market emerging, this is the phase where the big market delusion, a term I coined over a decade ago, is likely to emerge. That delusion has its roots in selection bias, where the people building products for the change to come tend not only to be true believers but also over confident, resulting in a collective over reach by companies and investors pricing these companies, and a correction.

Thus, bubbles are a feature, not a bug, when revolutionary change is a possibility. Finally, if business and investing is a combination of (business) stories with numbers, at this stage of the cycle, where there is little material that has already been accomplished, it is the story that drives growth and investment. Investors with actuarial or accounting mindsets will undoubtedly find these narratives unpersuasive and quickly consign these companies to the overvalued heap. While that impulse is entirely understandable, it is worth remembering that there will be other investors, who are willing to bet on optionality, where they invest in this space, hoping that the entities that they invest in will be the big winners (though they have not won anything yet) in a big market (which does not exist right now).

3. Business Building

Not all change is revolutionary, not all revolutionary change translates into big markets, and not all big markets create valuable businesses, and it is in the phase of business building that the truth starts to emerge. Since this is very much a test of businesses growing up, it represents a bar mitzvah moment for these businesses, in the sense that investors are no longer willing to just price on promise, and start demanding tangible evidence of progress. It is during the business building phase that you start to create the structure of converting products into businesses, with production processes, supply chains, marketing and distribution all taking form. In the process of building business, you will confront the realities that will determine whether you are a mass market or niche company, including unit economics and economies of scale. In the process, they will also discover a harsh truth, which is that many creative and talented product-builders lacking business-building capabilities, and either have to partner with someone who does, sell their products to established companies that already have systems in place (expect acquisitions, partnerships) or get pushed out of their own firms by their capital providers (venture capitalists).

4. Recalibration

As businesses start to succeed, the laws of economics, immutable and powerful, kick in. You should expect to see turnover and consolidation, as new entrants and existing players jockey for position, and business economics determine industry structure from splintered to consolidated to winner-take-all all possibilities. At the same time, the dark side also plays out as those (businesses and individuals) rendered obsolete by the change come under pressure, with some shrinking, some disappearing and some in denial.

It is worth recognizing that there is no steady state, because as businesses recalibrate to the innovation, they now represent the status quo and become targets for the next revolutionary change, Schumpeter’s creative destruction in motion. The picture below summarizes the cycle, mapping out the pathway from hype and hope to investing to harvest to building businesses to recalibration:

Where does AI fall in this cycle? ChatGPT, as I noted at the start of this post, may have been low-tech AI, but it got the hype cycle rolling, and social media amplified and accelerated that rollout, and its broad reach meant that almost everyone has seen it at work. The hope that AI’s popularity and reach would create a big market built up in parallel, with capital flowing into firms in its sphere (as well as wannabes that latched on to it, as a buzzword), pushing the market capitalizations of the companies building AI’s architecture (computer chips, LLMs, power and water companies, cloud) into the trillions of dollars, funded with equity and debt. It was not just financial market participants that saw its allure, as hyper scalers and new entrants invested hundreds of billions into AI cap ex, partly because they believed in its promise, but partly out of a fear of missing out. The graph below looks at cap ex in just six of the largest players in the space, four of them in the Mag Seven (Alphabet, Amazon, Met and Microsoft) and two outside (Oracle and Coreweave):

Source: Cap IQ

Cumulatively, the total investment from just these companies amount to $1.7 trillion, over the last few years, and their guidance suggests that they are not done, with trillions of dollars in AI cap ex commitments in the next three to four years. You can see why I use the analogy of a factory, and argue that AI has built the most expensive factory in history, and done so in hyper speed.

To what end? It is only in the last year or so that you are seeing the beginnings of business building, where companies are generating revenues from selling products made by the AI factory, with Anthropic and OpenAI as the most prominent examples. Those revenues are small for the moment, relative to capital invested, and the profitability is still a reach, but there is a host of experimentation going on on model type (open versus closed), business models (subscription versus usage) and pricing. The seeds of disruption have been sown, and there are signs that AI’s rise will make a significant dent in the profitability of some businesses, with technology companies in the software and intermediary segments being the first casualties. The AI story is clearly further advanced than it was a year ago, but it is still early, and there will be changes and challenges that face both the players in the space and the investors in these players, making this AI’s bar mitzvah moment.

A Business Framework for AI

If you are an onlooker or undecided on the AI question, I don’t blame you, if you find yourself whipsawed by what seem like persuasive arguments on both sides and waylaid by distractions aplenty. That is because there are so many strands to this story that taking any strand in isolation can lead you to a conclusion about AI as a business that is hopelessly of course. The best way to bring all these strands together is by going back to basics, and establishing the drivers of the value of any business (not just AI):

In this structure, there are three broad drivers that will determine how AI as a business will unfold. The first is with an assessment of the size of the total market for AI products and services, the second is the industry economics in that market, which, in turn, will determine how many companies will cater to this market and the profitability of these companies, and the third will be an assessment, or at least a preliminary judgment, on what the moats or competitive advantages will be in this business.

1. Market Size

It is true that the value of a business is tied to how big a market there is for its products and services, but it is also true that this metric, converted into an acronym (TAM) has become a gaming tool in the hands of founders, venture capitalists and bankers. In my SpaceX valuation, where xAI is the primary AI business, I noted that bankers estimated a total addressable market (TAM) of $22 trillion for xAI, which I felt was more hallucination than estimate. That said, any discussion of AI as a business has to start with the total market question, and it is worth starting that discussion with an examination of where we are right now in terms of revenues from AI products and services.

As we head towards the end of the third quarter of 2026, with chatter about Anthropic and OpenAI getting louder, both companies are racing to set up their stories by reporting their updated annualized revenue run rates (ARR), an admittedly self-serving metric (for growth businesses) estimated by taking the most recent period (week, month etc.) and extrapolating to a year. On August 17, Anthropic that its ARR at the end of July 2026 was $65 billion. OpenAI’s estimate of its ARR at the end of July was about $40 billion. While both numbers represented quantum leaps from their values just a year ago, adding these estimates to the revenues that SpaceX (from xAI), Microsoft (from its AI offerings) and other players, even with the most generous estimates, generate from selling AI products and services yields a total revenues that is modest:

As you look at these numbers, there are a few truths that are undeniable. The first is that the AI product and service market is not only fast growing, as evidenced by the ARR for the lead LLMs, but unpredictable, with Anthropic’s most updated ARR coming in $10-$15 billion below estimates. The second is that even with the most upbeat and optimistic estimates of revenues for AI products and services, the current revenue number caps out at about $250 billion, and that sounds like a big number, until you scale it to the trillions invested in the space. Put simply, the big winners in terms of revenues and operating profits, at least so far in this AI cycle, have been the companies that supply the infrastructure components, with chips (Nvidia) electrical equipment providers and power utilities all sharing in the spoils.

That, of course, is just the existing market and with immense growth built into it, the question becomes about the end game, and that end market size, at least at the moment, seems to be anyone’s guess. While the xAI bankers will undoubtedly use this uncertainty as a shield to not have to justify their estimate, there are ways we can start framing our choices, beginning with aggregate measures of what businesses spend as operating expenses, since AI’s big sales pitch is that it will lower that spending. In 2025, the aggregate operating expenses at publicly traded companies was about $65 trillion, broken down by sector and geography below:

Source: S&P Cap IQ

These expenses include the costs of raw material and inputs that are immune from AI’s efficiency push, since AI cannot replace the rubber you need to make tires, the wheat you need to process to get cereal and the chemicals that go into fertilizer. Consequently, it is the portion of these expenses that took the form of employee compensation, in all of its forms (wages, salaries, bonuses, stock-based compensation) that AI is targeting. While some companies break this portion of expense out explicitly in their financial statements, others do not, but there is macro data on this metric, albeit splintered geographically. Drawing on Federal Reserve data of compensation for all US employees, I get the following numbers:

Source: Federal Reserve

I would argue that AI’s total addressable market, in the US, cannot be greater than $12.96 trillion, the total employee compensation in 2025, or an inflation-adjusted variant, if it is in the future, it is roughly twice that amount, if you target global spending on employees. While that number is large enough to set AI optimists’ hearts aflutter, a world where every employee is replaced by an AI agent would not just be dystopian, but also an economic basket case. In fact, AI’s target market will be smaller, depending on the answers to four questions:

1. Tool or employee replacement: As AI products have become more powerful, the debate about whether AI’s future lies primarily as a tool or as replacement for human labor has also raged. In a post earlier this year, I focused on a Citrini report that played out the effects of the latter, highlighting the costs to the economy of laid-off white collar workers (and their income) and the effects on the market. In that post, I did note that notwithstanding public stories of layoffs in software companies, there has been little evidence (so far) of aggregate displacement of labor in any sector, at least so far. The takeaway, at least from this discussion, is that AI’s disruptor role will be far greater, as will its total addressable market, if it replaces employees, rather than is used as a tool.

The logic for why AI tools than AI as employee replacement will have a smaller market is simple one, from a business perspectives. Businesses spend money on tools, but that spending will be in addition to what they already spend on employees, and while they rationalize that spending with (promised) improved productivity, it has to be a fraction of employee compensation. You can also why the current AI players (OpenAI, Anthropic) are opting for speedier disruption over a slower one, because it will then increase their odds of winning, albeit with higher displacement costs for society.

2. Pricing of AI products: In the last year. Anthropic and OpenAI have garnered publicity for their most powerful products (Claude, Codex etc.), and while some of them do offer the capabilities that will allow them to replace workers, they are expensive enough that it will make sense to use them only for high-paid labor. In 2023, the US government estimated, based upon personal income statistics, that the highest quintile accounted for 51% of all employee compensation.

Source: Library of Congress

Bringing this factor into play with the total employee compensation of $12.96 trillion in the United States, you can argue that only about half of that market (at the most) is open to disruption from AI replacement products,

3. Breadth of use: There are some industries where AI will make more inroads, and do so sooner, than others, and what separates them will be the nature of work in the business. As we noted just a little bit earlier, software and coding have been the easiest entry points for AI products, since the output tends to be more rule driven and easily verifiable An article in the Harvard Business Review, for instance, measured the risk of displacement across different occupations:

Research from Suraj Srinivasan, Harvard Business Review

I would expect AI to be more successful, both as a tool and employee displacer, in settings where there is less client or personal interaction and more rule-driven than principle-driven jobs. If you look back at operating expenses, broken down by sector, the sectors most exposed to AI disruption (technology and financials) have aggregated operating expenses the amount to less than 20% of the global total, whereas sectors more immune (industrials, materials, real estate utilities) amount to a third of the global total.

4. Geography: If you consider the fact that AI is more likely to displace workers and generate revenues in non-manufacturing companies that have high priced labor, it follows that the disruptive effects of AI will be greatest in the United States and have a smaller footprint elsewhere in the world.

Source: Visual Capitalist

Looking back at the table that breaks down operating expenses geographically, for publicly traded firms, you can see that the US, Europe and China are the three biggest markets for AI, since these are regions of the world where companies spent most on operations in the aggregate

If you consider $26 trillion as your upper limit for AI’s total market, in current dollars, your estimate of the size of the AI market will depend on your assessments of whether you fall on each of the four factors, with the largest assessments of TAM emerging from a view of AI as an employee replacement that cuts across industries and geographies, but with a cost for AI agents low enough to replace workers with lower income. At the other extreme, your assessment of the TAM will be much lower, if you view it as a tool, no matter how powerful, with application in select industries and geographies.

Rather than view different assessments of AI TAM as “he said, she said” disagreements, the debate would be much more grounded if these assessors were explicit about where they stand on the dimensions (AI as tool or employee replacement, target high-priced workers or all employees, useful in a subset of industries or all industries and primarily US-based or global) that drive their estimates.

In my first economics class, I became very fond of the words "ceteris paribus", latin for "all other things being held equal", partly because it does make it easier to focus on what's changing, but mostly because latin makes you sound smarter than you really are. Holding all else constant may be reasonable, when change is small, but when the change is revolutionary, as AI could very well be, it is no longer reasonable to make that assumption. Consequently, I think it behooves anyone making an assessment of AI's total addressable market to think through the implications for the rest of the economy and the world, because there may be constraints there that need to be considered. The most optimistic scenarios for AI investors and businesses, where the market for AI products is largest, are also scenarios where the damage done to other businesses is greatest and the impact on employment and the economy, in the near term, is most negative. Put simply, if the AI market explodes in size and does so quickly, and millions of high-paid workers lose their jobs, the economic damage will be deep, and there may very well be insufficient income to buy the products that the AI revolution creates. It is possible that in the long term, AI's benefits may exceed its costs, but that will occur in a very different economic setting than the one we have right now.

Every revolutionary change gives rise to new businesses, some of which are extraordinarily successful. The internet business was the starting point for online retailing, with Amazon as its most successful player, streaming entertainment, with Netflix and Spotify, as players, and search, with Yahoo! and Google cashing in. The smartphone was the vehicle used by a host by intermediary businesses, with Uber, Airbnb and Doordash, all changing the businesses they entered, as well as social media, with Facebook and Tiktok emerging. If AI’s promise plays out, it is almost certain that it too will be the launching pad for new businesses. If you are tempted to count the revenues of these businesses as part of AI’s payoff, there are two reasons for caution:

1. Net benefit test: All of the companies that I mentioned in the last paragraph were immensely successful, but some or much of their success came at the expense of existing players. Notwithstanding Amazon’s astonishing success in online retailing in the last 25 years, retailing as a business has not seen explosive growth, since brick and mortar retailers have largely been left in the dust. The success of entertainment streaming companies has not only come at the expense of traditional entertainment companies, but in the aggregate, it has made the business less profitable. It is true that ride sharing has caused the car service business to become larger, as ride share use changes user behavior, but it has contributed to wounding the auto business. The most telling statistic on the impact of disruption, especially from technology companies in the last four decades, is that real GDP growth has been mostly stagnant in much of the world for that period, suggesting that the net impact on economic activity is modest, at best.

2. Sharing the winnings: Even if you take the gross added value from new businesses (in revenues, earnings or value) for revolutionary change as a given, it is unclear what the companies that were the architects of that change gain as a result. The internet was built by telecom companies (with investments in cables and phone wires and supplemented by equipment companies (like Cisco), but none of them benefited from the rise of internet companies like Amazon. The smartphone was built and popularized by Apple and Samsung, but they don’t get a slice of what Uber and Doordash make, when customers use smartphones to access their services. Apple benefits at the margin, as customers are coaxed into a cycle of replacing existing phones with newer versions, but the benefits are tangential.

If you look at the history of revolutionary changes, from the steam engine (factory age) and railroads, all the way to the technological changes of the last four decades, it has generally been the case that while the companies that build the architecture for these changes initially benefit, the big winners often catapult themselves off that architecture. You can make the argument that AI will be different, but for AI architecture companies to gain a slice of the new businesses that emerge from their technologies, they will need different business models.

2. Industry Economics

Accepting the premise that AI products and services will have a big market, in terms of revenues, is just the first step in structuring its business argument, and in this section, I will focus on converting those revenues into profitability, by first looking at the business models that AI producers can consider adopting, with pluses and minuses, as well as unit economics and economies of scale, measuring what it will cost companies to produce the AI products that they are selling.

a. Business Models

In keeping with the trial-and-error process that characterizes young industries seeking workable business models, we have seen AI businesses experiment with different versions on at least two dimensions:

• Subscription versus Usage Models: Early in AI’s business evolution, subscription models for both individuals and enterprises were the entry point for AI companies, and while those subscription models still bring in revenues, companies have shifted away from them for a simple reason. Unlike businesses like streaming or even software, where the marginal cost of an additional subscriber is zero or close to zero, AI products and services are expensive to generate, in terms of compute costs and data, and both Anthropic and OpenAI have discovered that subscribers, left unchecked, quickly become cost generators rather than profit centers. Thus, it should come as no surprise that Anthropic has shifted to usage-based models, where users (especially at the enterprise level) pay based on how much and how intensively they use AI products. OpenAI is still more dependent on subscription models that Anthropic, but it too is seeing a shift to usage-based models.

  

Source: Leaks and estimates

While advances may change the economics, it looks like while AI-subscription models will stay available, they will come with strict limits on usage, and that the bulk of revenues in this market will be based on usage. At the same time, there will be differences across AI companies, based on whether they are targeting the premium AI market, where usage-based models will dominate, or the mass market, where subscription models will continue to be offered.

• Open versus Closed Models: There is an open and vigorous debate going on in AI circles as to whether the AI businesses should offer clients and customers open models (where clients can modify, adapt and build on the models) or closed models (where they are not allowed to do so). This debate is complicated because there are multiple forces that come into play including how power in this space is concentrated (with closed models giving its makers more power), how much privacy they offer (where the argument is that open models require less sharing of private data) and how safe they are (where it is posited that open models can be more easily hacked and turned into disinformation). All that said, there is clearly a business economics twist to this debate. Closed models give the companies that sell them more pricing power (higher margins), and perhaps are stickier (making it difficult to switch away), because they are customized, but they require more resources to build and maintain (higher costs) and may work only with premium products. As with subscription versus usage models, you are likely to see divergence, with companies targeting the premium market more likely to stay with closed models and those building more workhorse applications trying their hand at open models.

As I noted at the start, it is still early in the game, and as technology, regulation and cost structures shift, not only will there be more twists and turns involved, but it is likely that in steady state, we will have different choices for different segments of the AI markets, more subscription-based and open models in mass markets and more usage-based and closed models in premium markets.

b. Unit Economics

Through much of its existence, technology, as a business, has largely benefited from beneficial unit economics. It costs a software company almost nothing to sell its next unit, and for most platform companies (Netflix, Uber, Airbnb), the cost of adding a user or subscriber, once the platform is constructed is negligible. As a consequence, being the largest player or a first mover can put you on a pathway to industry dominance, with large market share and high profit margins thrown in. Early in its life, AI was thrown in by some into the technology pile, and it was assumed that it too would have the same characteristics, i.e., that it would cost little or nothing to produce the additional unit and that high margins and dominant market shares will follow.

With the caveat that things can change quickly, the evolution of the AI product and service market has turned out to be different, and some of those differences look like they are baked in. Beginning with the fact that AI products require more capital intensity, in terms of data center build and cap ex prior to operation, they are also proving costly to produce, at least in their most powerful forms. One way to see the dual forces driving unit economics in AI at play is to focus in on AI tokens, the currency of AI production. The good news, in terms of unit economics, is that the cost of producing a token has dropped dramatically, as AI infrastructure gets built out, but the bad news is that the tokens used to create AI products has surged almost as dramatically, as these products become more powerful. As a result, the price of accessing frontier AI models has increased over time:

Sources: Token Price Index

These two trend lines point to a divergence that is coming to the AI products and services market. In mass market AI, where AI tools are basic and don’t need power boosts, you should expect costs for AI products to go down. In the premium market, where you are building AI products either as tools on high-end tasks or as replacement for highly priced labor, the costs will be tougher to reduce, if each upgrade in product power puts demands on the inputs - more data to process, more power to run data centers and more powerful chips in the data centers - that causes these input costs to rise.

c. Moats and Competitive Advantages

Allowing unit economics and economies of scale to play out in the AI business, you still have a final piece of the business puzzle to consider to make the leap to profitability, where you bring, as you would in any business, the moats and competitive advantages that will separate the winners from the wannabes. Since this is a question that we ask about every business, it makes sense to start this discussion by looking at potential competitive advantages in any business:

Focusing in on AI, the question that we first face is which of these moats is most likely to be the defining one in AI, and I believe that the answer depends on which segment of the AI market you are looking at:

• In mass market AI, where products and services are standardized and basic, you should expect competitive advantages to flow from unit costs being lower at a company than at its competitors, either because of scale (with bigger companies having an advantage) or proprietary access to data.

• In premium AI, where products are customized, powerful and pricey, you should expect the winners to be companies that have the technological know-how to craft these products, while bringing the costs of delivering power under control. In addition, products that are built around client data, especially if the client is protective of that data, will become stickier and more difficult to displace, giving companies that make them more pricing power for a longer period.

Does having the people perceived as the smartest in the AI space working for you give you a competitive advantage? At the moment, the answer seems to be yes, and you saw this phenomenon at play when Jeff Dean, chief scientist at Google DeepMind (its AI entity), left the company in 2026 to create his own AI start-up, and the market reacted by knocking Alphabet’s market cap down by 5.4% (more than $100 billion). In the same vein, the AI firms (especially Anthropic and OpenAI) have been raiding universities for their computer science and technology talent, with some AI-focused economists thrown into the mix and even a few philosophers. I think that the attention paid to these people hires and departures are indicative of how young this industry is, and how much its success will depend on building products right and marketing them to the right customers, and as it matures, I expect this factor to fade in prominence.

Is there a potential brand name advantage? Put simply, would you be willing to pay a premium price for a Claude agent over an AI agent crafted by a different company? The answer is still being worked out, because in addition to all of the business elements of this choice (product power, reliability), trust is a factor, since client companies are giving AI products access to secrets and data. It should come as no surprise then that AI companies are all competing in the virtue space, where each one puts itself out as more trustworthy and caring about public good than the next one. It may be cynical of me, but when I hear Dario Amodei or Sam Altman wax eloquent about how they plan to protect the world from the dark side of AI, I feel the urge to quote Shakespeare, and say “thou doth protest too much”. Ultimately, actions speak louder than words, and these companies will be judged based more on how they behave, when confronted with ethical questions, than on what the write about themselves.

3. Constraints and Limits

While much of this post has been about AI’s business prospects and evolution, there are parallel discussions that are occurring about AI’s impact on society. There are four reasons why AI’s social and cultural effects are being so widely debated:

1. Real estate footprint and resource usage: AI, in terms of the investment footprint it is creating, is closer to the railroads in their early years than it is to any technology company. Like railroads, AI requires data centers that sprawl over huge areas, and unlike railroads, many of these areas have people living in them, whose lives will be altered by the presence of these centers. While the proponents of data centers have sold them on the basis of the economic benefits they bring, and these benefits can be real, it is quite clear that for many people who live in the vicinity, the upending of their lives is not worth the benefits. As a result, the backlash against data centers is real, showing up in politics at not only the local level, but also at the national level; it is quite clear that at least in the United States, it will be a lead topic, perhaps even a wedge issue, in the next presidential election. Another reason that data centers lead to resentment is their disproportionate use of power and water, and even if that cost is pushed back to AI companies, the costs to the planet are still being totaled.

2. Data privacy and power: Earlier in this post, I pointed to privacy as one of the dividing lines in the choice between open and closed AI models, but that is a small part of a broader question, which is about data being accumulated by and mined at AI companies. After two decades of seeing social media companies step across the privacy line in their use of private data, it is understandable that there is wariness about granting access to even vaster amounts of data to the likes of Anthropic and OpenAI.

3. People displacement: Going back to the discussion of the total market for AI, I noted that the best case scenarios for AI, i.e., the scenarios where the market for AI products and services will be greatest, are also scenarios where there will be not just be significant job loss, but losses in high-paying jobs. While tech advocates will point to new jobs that will be created to replace the ones lost, that transition, even if it happens, takes time and will come with pain. Looking back at the disruption of blue collar jobs in the US and Europe, from China, that disruption created pain, albeit unevenly shared, and has led to political and economic aftershocks that are still playing out across the world. If AI’s disruption plays out on a broad front, the resulting displacement will be much larger, in terms of economic impact, and perhaps much more painful.

4. Income equality and fairness: For much of this century, one of the recurring themes in both political and economic circles has been the growth of the divide between the super rich and the rest of society. There is a suspicion that AI will make this divide wider, and that suspicion will only intensify with each AI company that goes public. SpaceX, when it went public at a market capitalization of close to $2 trillion created a host of centi-millionaires (worth more than $100 million) and the same phenomenon will play out when Anthropic and OpenAI hit the market. I am not a fan of setting economic policy based on greed and envy, but you can see the pushback against inequality playing out in the political arena.

If you are tempted to ignore these discussions, because you are an investor or interested only in the business aspects of AI, it is only a matter of time before these spill over into economic consequences, and then into the metrics that drive business value.

1. Data centers will get more expensive to build, and power and water will be more tightly rationed, leading to companies having to spend both more on their upfront and capital expenditures, and as costs in generating AI products.

2. As concerns about data mount, there will inevitably be scandals around the misuse of data, and those scandals will lead, as they did at social media companies, to tighter restrictions on the use of data and higher costs is acquiring and protecting that data.

3. The worries about being replaced by AI agents will create counter movements, starting with system requirements that preserve jobs for humans (even if AI makes what they do obsolete) and in some countries, requirements that their employers continue to pay displaced employees for extended periods, in the event of layoffs.

4. If AI creates its own cache of billionaires and centi-millionaires, the push towards wealth taxes, no matter what you may think about their effectiveness or fairness, will intensify, as will the creation of new tiers in the tax table for higher incomes, and perhaps for AI income.

It has always been difficult to start and build businesses in new industries, and ti becomes doubly so when the rest of the world consigns you to the dark side. If you are an investor in this space, especially one excited about the size of the potential market and pathways to profitability, you need to be realistic in incorporating the constraints that are already cropping up, and will become more binding, into your valuation.

From Macro to Micro: Zeroing in on company valuations and investments

This post has focused on AI as a business, and if your interest is in valuing an enterprise in the space, whether publicly traded like SpaceX or Alphabet or privately owned (but heading for an IPO like Anthropic or OpenAI), you may wonder how it helps you in that endeavor. There are two ways you can approach these valuations. In the first, you can start with the market capitalization (actual in the case of publicly traded companies and estimated in the case of private businesses) and examine what you would need in terms of end revenues and profit margins to justify the market capitalization. In the second, you can start with the company that you are valuing, and lay out a roadmap for that company in terms of product choices (premium or mass market, open or closed) and estimate what it will be able to deliver in terms of revenues, profits and cashflows over its lifetime.

1. Reverse Engineering Breakeven Points

In a post, late last year, I looked at the handful of companies, mostly tech, that have trillion dollar market capitalizations, and rather than take the knee-jerk reactions, i.e., that they must be overvalued with that market cap or that they must be great companies, because the market assigns them a lofty price, I looked at what these companies will have to deliver in terms of operating metrics in the future to justify their valuation.

We are effectively reversing the intrinsic value process, and trying to answer the question of how much revenues will have to be in a future year (where you specify when the company or business will be mature or steady state), given your company characteristics in terms of risk, profit margins and reinvestment needs. With Anthropic, for instance, where the rumored pricing for the IPO is $2 trillion, allowing the company premium pricing margins (after-tax operating margin of 30%) and above-average risk (cost of capital of 10%), the company will have to generate close to $1.2 trillion in revenues, if the AI market matures in ten years, and close to $2 trillion, if the wait is 15 years. That should give ammunition to both those bullish about the company, because in their story line, Anthropic products will be premium priced and replace workers across industries and geographies, and to those who are bearish, since that scenario looks unlikely.

In fact, you can consider AI companies in the aggregate, by adding up the market capitalizations of companies that are already public (or at least the portion of the revenues that come from AI) as well as the rumored pricing of companies like OpenAI and Anthropic, waiting to go public, and going through the same exercise. Using an aggregated market pricing of $5 trillion (probably a conservative judgment, given the VC pricing of hundreds of companies in the space) attached to all AI product and service companies, and assigning a blended operating margin of 20% for the industry, the revenues that you would need for the entire business to breakeven would be $5 trillion, with a 10-year wait, and more than $8 trillion, if the wait is 15 years. Looking back at the discussion of the total addressable market in the earlier section, you can see that this would represent quite a reach, a manifestation of the big market delusion.

I will be the first to admit the limitations of this reverse engineering, but when data is still scarce or non-existent, it does provide a framework for reasonableness and a constraint on story telling. In fact, you can use this framework to examine what any investment, whether it be the price you pay as an investor for an AI business or the capital expenditure into AI made by a company, will have to generate to break even as an investment. Thus, if you are questioning whether Microsoft or Meta’s AI cap ex is value creating or destroying, you can use this generic breakeven spreadsheet, to make your own judgment.

2. Build up to value

If you follow the AI business script laid out in the class, you also have a process for valuing any AI company that hopes to make money in this space, but to put this process into the play, here are some of the issues that you will have to address to estimate value.

1. Product choice and market focus: If the AI market splits into premium and mass-market product market, the first step in valuing any company in this space will be to make a judgment on which of these markets the company will target, and how much of its revenues will come each of the segments.

2. Unit economics and economies of scale at company: The choice of market segment matters because the unit economics and economies of scale you assume for the company will have to be consistent. AI companies that offer mass market products will charge lower prices, with a greater percentage of revenues coming from subscriptions, but will benefit more quickly from improving unit economics, as the cost of AI tokens continues to fall. In contrast, AI companies that target premium markets, will earn higher profit margins and have stronger moats, but struggle more with unit economics, as token usage increases with product power.

3. Competitive advantages and moats: The types of competitive advantages (moats) that the company you are valuing will seek out, and lock in, if successful, will also vary depending on the targeted market, with cost advantages and scale working in the company’s favor, with mass markets, and technological edges and product stickiness being more sought after, with premium products.

4. Investment needed to deliver growth: While AI companies are more capital intensive than their tech counterparts, the additional reinvestment needed to deliver value can be altered by investments already made by a company. Companies that have built capacity in advance of growth will be worth more than companies that will have to reinvest contemporaneously to deliver growth, and companies that find ways to invest more efficiently will also have higher value. It is interesting that starting with Deepseek, China seems to be trying the latter path to AI dominance, using less expensive (and less powerful) AI chips and not investing as much in mega data centers, and it may very well be the right choice, for much of the AI product and service market.

5. Regulatory constraints (current and in the future): To the extent that regulators and governments have a great deal at stake, the value of a company can be affected by where it operates geographically and the rules and regulations that govern AI products in that geography. If past behavior is an indicator, the EU will be an inhospitable setting, for AI products and companies, and that may make a difference in how you value Mistral, with a base in France and more European-focused clients.

I did try my hand at this process, when I valued xAI as part of SpaceX, and I learned from doing so, but the company’s stakes in space launch and internet service did muddy the waters. As Anthropic and OpenAI move towards their public offerings, I am looking forward to applying the framework developed in this post to those firms, when their prospectuses are made public. In keeping with my belief that it is best to be open about biases, I will confess that the rumored pricing for both companies ($1.5 to $ 2 trillion) looks rich, but I am open to being surprised.

Conclusion

At the start of this post, I noted that I was writing this post for myself, because like many in this space, I was finding myself pulled in a dozen different directions on AI, and desperately in need of a framework for thinking about whether I should be paying attention in the first place, how to reconcile competing viewpoints and what it means to me, as an investor. This is my try at creating a comprehensive framework, and I am sure that there are elements that I have missed and holes in my thinking, but it is a start. I am clear eyed about what this AI framework will not and will do for me; it will not tell me what the TAM for AI products and services will be or whether Anthropic is worth $ 2 trillion, but it will give me bounds for my estimates of TAM, allow me to determine that a $22 trillion TAM for AI is fiction and recognize that having your ARR grow 80% a year last year is not even close to being a rationale for why you should buy Anthropic at a $2 trillion pricing.

As you work your way through the framework, there will be room for significant disagreement on the reach of AI and its value as a business and anyone who claims to have conviction that they know what’s coming is being either ignorant or arrogant. That ties in well with my last post, where I examined the swift rise and the even swifter fall of Leo Aschenbrenner, whose entire investment strategy was built around the conviction that AI would decisively and quickly win the disruption war. The problem that I noted was not that his vision was not plausible (it was), but that it was definitely not certain, or at least assured enough to borrow immense amounts to fund it.

YouTube Video

Data

1. Employee compensation for all US workers - 1929 to 2025

2. Revenues & Operating expenses, broken down by sector and geography

Spreadsheets

1. Generic breakeven calculator (for both company pricing and AI cap ex)

2. Anthropic breakeven

3. AI Business (aggregate) breakeven

The reads on the swift rise and fall of Leo have been fascinating, a Rorschach test of investing priors. For older investors, the lesson was that you can be blessed with intelligence, but that wisdom required experience, which, at least in their saying, conveniently comes with age. For value investors, many of whom chafed at Leo being hailed as the next Buffett, there was vindication that there will never be another Buffett. For AI skeptics, who have long been on the lookout for catalysts that will break AI fever, there was at least a brief moment of hope that this was the catalyst. There is some truth and some overreach in each of these responses, and I don’t plan to rehash them. Instead, I would like to use this story to talk about investing conviction, words used mostly in a favorable way, when people talk about success in markets. I am not as convinced that investing conviction is a net plus, but to get to that conclusion, I think we need to look at what it is, where it comes from and what it leads investors to do.

The Story of Situational Awareness

The Situational Awareness story is inextricably tied to the story of Leopold (Leo) Aschenbrenner. The short version of his life story is that he was born in Germany in 2001, and enrolled at Columbia University when he was 15. After graduating with a degree in economics in 2021, doing research briefly at Oxford University and working at Sam Bankman-Fried’s FTX fund, Leo joined OpenAI as part of the team working on AI safety. He was fired in 2024 for leaking data on the firm, though his motivations for doing so are murky and he contests the allegation, and he published a long (167 page) paper titled “Situational Awareness: The Decade Ahead”, which was not only widely circulated, but also became the blueprint for the fund that he created.

The Situational Awareness fund, founded in July 2024, and initially funded by tech luminaries, quickly took off as its bets on AI chips and infrastructure and against AI-damaged sectors paid off. Its early success allowed it to attract more money, with Jane Street being one of the more prominent names involved., and with the additional capital in play, it expanded its presence. While most of the companies that made the fund’s list were publicly traded, it also had a stake that Leo had acquired in Anthropic, still a private business. The numbers posted by the fund made investors notice, as can be seen by its rise From August 7, 2025, to June 23, 2026, its highest mark day:

Source: Portfolios Lab

Measuring returns from August 7, 2025, Situational Awareness was up about 367% through June 19, 2026 and its returns since founding are even more stratospheric. Even at its peak, there were three caveats that any investor with experience in the market would (or should) have brought up. First, in market time, where decades of over performance are needed to separate luck than skill, a fund that has been successful for a little more than two years qualifies more as a shooting star than as a beacon of light. Second, to deliver returns of this magnitude, you not only have to be right in your guesses, but those guesses must be super-charged by adding substantial leverage to your strategy, either explicitly (by borrowing) or implicitly (by using options). Third, the fund followed the classic 2% (of funds under management) and 20 (% of specified upside) fee structure, an abomination that not only creates an almost insurmountable handicap, in the long term, for investors in the fund, but also encourages reckless risk taking on the part of management.

If the rise of the fund was breathtaking, its fall was even more so, and you can see the meltdown in the four weeks of July in the graph below, which looks at the fund performance from June 19, 2026 to July 29, 2026, the last day of trading for the day, before the fund was liquidated:

Source: Portfolios Lab

Note that the loss of principal (of more than 43%), with the Anthropic holding value retaining mostly intact, as a private holding, but with the public investment portion of the portfolio down by almost 67%. I am sure that there will be case studies and forensic analysis of what happened in these weeks, but for me, the lesson is one of market symmetry. What the market gives easily, it also takes away just as easily, and if you put into place strategies that are designed to deliver outsized returns, you have to live with the reality that you can have outsized losses. The surprise, though, for many is not that the fund lost money in July, but that it did not survive the month, and was forced to sell much of its public investment portfolio to Citadel, at prices, that at least in hindsight, look like bargain basement levels.

Conviction: What is it and where does it come from?

How do I get from the Situational Awareness story to a discussion of investment convictions? Simple! Leo’s core belief that AI would win the battle with the status quo in most businesses, and that the win would be decisive and quick, was not unique, and not only are there other investors who shared that view, but there are also companies that are investing in AI cap ex, driven by that view. That said, to get from that view to a hedge fund built entirely around buying AI winners and selling AI losers requires that the belief to be deeply set and using debt to magnify those returns suggests strong conviction.

What is investment conviction?

Before we embark on a discussion of investment conviction, it is worthwhile to start with an understanding of what it means. While there a myriad of definitions out there, the general consensus is that investment conviction measures the belief that an investment opportunity will generate significant returns, relative its risks, and with that definition, you can see that conviction is a continuum, rather than an absolute.

At one end of the spectrum, you have absolute conviction, where you know (or think you know) with certainty that an investment will pay off. At the other end of the spectrum is investment mush, where your feelings about an investment paying off are so weak that you are unwilling to even voice that opinion, let alone put money behind it. With most investments, you fall in the middle, with the variations being in the degree of confidence that you have in being right.

As you can see, the elevation of conviction as something to be sought after in investing is because in the complete absence of conviction, you will not act and that paralysis can result in portfolios entirely or almost entirely held as cash. I don’t think, though, that even the strongest proponents of conviction as a good quality in investing would make the argument that you should feel certain about the outcome of investments, when uncertainty is part and parcel of investing.

Where does conviction come from?

So, what is it that determines investment conviction or the lack of it? To generate a basis for that discussion, let’s go back to basics, and start with an assessment of what has to happen for an investment to be viewed as a money maker. No matter what your investment philosophy, the process starts with a market price for an investment, and an assessment of what you believe is a “fair price” for that same investment. I am being agnostic about how you arrive at the fair price, leaving the door open for chartists, who may find it by looking at past price patterns, fundamentalists, who believe that you can assess fair price, only by looking at the fundamentals of the investment and traders, who may be in possession of information that leads them to believe that the current price is wrong. For the process to deliver winnings, though, there is a second part to the process that often gets less attention, which is that the market has to correct, with the market price moving towards or even to your fair price.

To have conviction in an investment, you therefore need to believe strongly in three aspects of this process:

1. That your assessment of “fair” price is right (or at least more right than the market consensus)

2. That the market will correct, i.e., that the market price will move to, or towards, your fair price

3. That this correction will happen during the time you plan to hold the investment (your time horizon), either because the investment has an expiration date (maturity) or because you feel that there will be a catalyst that causes the correction.

With this description in place, you can see that investment conviction will depend on the investment in question, the market that it is priced in and on the investor making the judgment.

a. Investment Mispricing

If the investment process starts with an assessment of a fair price that is different from the market price, there are at least four reasons why you may be more confident in your assessment of the price of an investment, relative to the market consensus:

1. Private information: At the risk of treading on or crossing the line between the legal and the illegal, you may be in possession of information about an investment that is not available (or at least widely enough available to the public to be priced in) that you believe will change its price.

2. Information processing: To the extent that private information is rarely available to investors, and even if available, difficult to act on legally, much of active investing is built around collecting and processing information that is available to the market. That private information can range from past prices and trading volume (used by technical analysts) to public filings (the financial data that the company provides, often the basis for fundamental investors) to quasi-public information (analyst forecasts and revisions, hedge fund and mutual fund holdings). If you are using this information to assess a fair price, it is because you believe that you have found patterns in the data that others have not.

3. Business understanding: In some cases, your fair price will derive from your belief that you understand the business economics for a company better than other public investors do, with your superior understanding coming either from working in the business or from technical training. This is especially true in complex businesses (like bio pharma) and complicated assets, and likely to be more the case with young companies, where financial history can stand in for business understanding.

4. Pricing mistake: There are some investment theses that start with a market mistake, whether it be in pricing an individual asset or a pair of assets. With a pair of assets, often with similar fundamentals, the mispricing manifests with one of the assets being mispriced against the other, and the correction take place when the mispricing disappears. It is at the heart of derivatives trading, where options or futures on a traded asset can be mispriced enough that you can lock in the profits from the pricing mistake, with the guarantee of correction, when the option or futures expire.

In sum, you can see that the degree of investment conviction that an investor has can vary across different asset markets (real estate, equities, fixed income, derivatives), geographies (developed versus emerging markets) and within equities, across industry groups and sectors (less for technology and more for utilities, for example).

b. Market correction

Investment conviction may start with the spotting of a market pricing mistake, but for conviction to build, you need to get a measure of when, why and how the market will correct its mistaket. Here are some of the forces that can cause variations on the market correction dimension:

1. Finite maturity versus indeterminate end game: An investment with a finite maturity date comes with a greater likelihood that prices will correct than one without. A bond that is mispriced, by itself, or against other bonds of equal maturity, comes with a greater chance of correction than a stock that is mispriced against its own fundamentals or a paired stock. For the same reasons, if you can lock in a mispriced option against its underlying asset (or stock) or against other options of equal maturity, you are moving the odds in favor of correction. It should come as no surprise that true arbitrage, i.e., positions that can lock in guaranteed profits that exceed the riskfree rate are almost always in the fixed income or derivatives markets and that much of what passes for arbitrage in equities is quasi or pseudo arbitrage, where risk remains in the position.

2. Market frictions: Market mispricing can sometimes reflect market inattention or irrational trading, but they more often are the consequence of market frictions, including restrictions on selling short and exerting control over an investment (such as buying and liquidating a company). If that is the case, it is possible that market mistakes, while visible and seeming obvious to everyone involved, may never get corrected, or at least not get corrected until the friction is removed.

3. Market liquidity and depth: In equity markets, where there is no date by which mispricing has to disappear, corrections often require catalysts, i.e., events that lead market participants to reassess a market price, and to correct it. Those catalysts can come from corporate disclosures (earnings reports, for instance) by the mispriced company, corporate restructuring (divestitures and spin offs) or high profile investors (activists taking a stake in a under priced company or selling short an overpriced one). All of these catalysts are more likely to be present in liquid and deep markets, where information flows are more frequent and varied.

4. Investment time horizon: No matter what the market mistake, it can be argued that having time as an ally, and being able to wait for longer periods, for market corrections, should not only give you a greater chance of gaining from a market correction, but also give you more conviction in your investment, other things being equal.

The bottom line is that you can feel (relatively) certain that an investment is mispriced, but have no conviction in that investment, if you have no sense or faith that the market will correct in your time frame for investing.

c. Investor characteristics

Is it possible for two investors to find the same market mistake, with the same underlying rationale for the mispricing, and have different degrees of conviction in following through? Absolutely, and while part of the reason is differing time horizons, there are other investor-specific forces that also come into play:

• Intelligence and educational background: This may be a generalization, and if you disagree, you should take exception, but the smarter an investor is, and the more exceptional his or her educational background (the right schools, credentials and certification), the greater the conviction that investor is likely to bring to investments. One reason is that it becomes easier to attribute perceived market mistakes to the lack of intelligence of other market participants than it is to take a closer look at real reasons why they may not be mistakes in the first place.

• Personality: Like conviction, self confidence falls on a spectrum with wide differences across human beings. Some investors measure low on the self confidence scale, and look to others for big decisions. Others measure higher on the self-confidence scale and are willing to make decisions with incomplete information and in the face of uncertainty and disagreement. Still others have so much self confidence and so little self doubt that they risk having convictions that are out of sync with reality. Over five decades of research in behavioral finance has highlighted overconfidence as one of the key drivers of irrational investor behavior, and underscored the reality that not only does this trait vary widely among individuals, but also that the most overconfident players often rise to the top of the investment and corporate world. In the conviction discussion, overconfidence enters the game early and is perhaps the best explainer of why some investors, looking at what they think is a market mistake, feel so much more convinced that they are right than other investors looking at exactly the same mistake.

• Track record: When you invest, you receive almost instantaneous and continuous feedback on whether your investments are paying off, and while investment success is always preferable to failure, it is undeniable that some of the worst investment lessons are learned from that success. Much as Wall Street likes its adages of “not mistaking dumb luck for skill”, investors are quick to forget them when an investment bet that they make pays off, especially when their success leads to iconic and admiring profiles (”the next Buffett”) and investor capital pouring in.

You will notice that I don’t list age as an additional characteristic, but that is because I don’t equate aging with wisdom or temperance. It is true that aging usually muddies your investment track record, and that there is no better bound on over confidence than losing lots of money on what you thought was a sure bet.

The Consequences of Conviction

At this point in the post, I don’t blame you for wondering why conviction is such a big deal, since buying Palantir or SpaceX with low conviction counts just as much buying shares in these companies with high conviction. Conviction matters in investing because it affects two choices that investors make - the sizing of positions (with more conviction leading to larger positions in the same investment) and the use of financial leverage (with more conviction often translating into a willingness to borrow more money to fund the investment).

Concentration

One of the fundamental questions in investing, and one that evokes strong disagreement, is how much, if at all, investors should spread their bets. The debate is an old one and there are many views that fall between two extremes. At one end is the advice that you get from a believer in efficient markets: be maximally diversified, across asset classes, and within each asset class, across as many assets you can hold. The proverbial “market portfolio” includes every traded asset in the market, held in proportion to its market value. At the other is the “go all in” investor, who believes that if you find a significantly undervalued company, you should put all or most of your money in that company, rather than dilute your upside potential by spreading your bets. The discussion of investment conviction ties into the answer to this question.

I am not an absolutist on this front, because what you do as an investor should reflect your circumstances. At one limit, if you are certain about your assessment of value for an asset and that the market price will adjust to that value within your time horizon, you should put all of your money in that investment. This may seem like an impossible dream, but it it is what you hope to pull off if you find mispricing in the bond or derivatives markets (true arbitrage), where you can lock in the mispricing, with a guaranteed price correction at maturity (of the bond or options). At the other limit, if you have doubts aplenty and no conviction in your investment choices, you should be as diversified as you can get, given transactions costs. If you have no transactions costs, you should own a little piece of everything, a very real choice in a world of index funds and ETFs. After all, you gain nothing by holding back on diversification and your portfolio will be deliver less return per unit of risk taken. If you are active investor who is constantly in this position (of having no conviction), it is best to retire the “active” part of your investment profile and go all in on index funds.

For most active investors then, the question of how much to diversify will depend in large part on how strong or weak their investment conviction is, and that, in turn, depends on what investments these investors are focused upon. In this context, I would argue that the right amount of conviction and by extension concentration (or diversification) will depend upon the types of companies you invest in (with more diversification needed when you invest in younger companies) and your time horizon (with concentration increasing with time horizon)

Leverage

Financial leverage is an instrument that investors can use to enhance returns, but its use in investing has always been controversial. While borrowing money to fund an investment will increase upside, if you are right, it will also magnify downside, and in some cases, precipitate collapse, if you are wrong. In the early year of investing, financial leverage generally took the form of borrowing money to buy riskier investments (stocks), but the wave of default and distress triggered for investors by the great depression led to restrictions on the use of margin in stock markets. However, those restrictions varied across investor groups (with individuals facing more restrictions than institutions) and across asset classes (with more leverage in real estate than in stocks). The growth of derivatives markets opened the door to bypassing these borrowing restrictions, since buying a (naked) call option is equivalent to borrowing the underlying asset with debt, with leverage increasing with how out of the money the call option is.

Since leverage magnifies both upside and downside, it stands to reason that investors with more conviction in their investment, i.e., that it is under priced and that correction is imminent or very likely to happen, will borrow more money than investors with less conviction:

When you are certain that an investment will pay off, you can use maximal leverage, and in true arbitrage, this leverage can be used to convert small mispricing into pure profits. In fact, when options and futures are mispriced relative to their underlying assets, you can borrow 100% of your investment needs, and since the pricing error will disappear at maturity, make a pure profit.

In sum, financial leverage magnifies your core investment judgments. If they are good, leverage will make them look better over time, and if they are bad, they will make them worse. That said, there is a component to the use of leverage that needs to be brought into the picture. Even if you are a good investor, with solid conviction in your investment ideas, using debt to turbocharge your returns can sometimes shrink your time horizon by forcing you to liquidate your mispriced investments before the market corrects its mistakes. That truncation risk eliminates the possibility that you can come back from your losses, and perhaps even have your investment thesis vindicated.

A Corporate Life Cycle Perspective on Conviction, Concentration and Leverage

I use the corporate life cycle construct in almost every aspect of finance, because it not only helps understand corporate and investor behavior, but also provides perspective on why one size (or proposal) does not fit all. The corporate life cycle maps out a firm’s evolution from a start-up to a growth company to maturity and eventual decline, and traces out changes in revenues, earnings and risk over the aging process:

There is no deep intellectual insight here, but as companies age, the challenges that investors face in pricing them shifts. With young firms, where there is little history, the business model is still in flux and the value lies almost entirely in the future, the pricing will be less precise than for more mature firms, where there is an established business model and more financial history, and more of the value comes from investments already made. For declining firms, where liquidation looms as a viable option, the pricing exercise becomes one of estimating liquidation value, i.e., what others will pay for the individual assets owned by the firm. The higher pricing uncertainty that investors face when valuing younger companies is accompanied by a second problem, which is that pricing mistakes, if they exist, need catalysts for market correction. Those catalysts are less frequent and decisive with younger companies, where earnings reports have little of substance to report and operating targets are diffuse, creating more open questions about whether a market mistake will be corrected, and if so, when that will happen.

Using this framework, you can see that you can be more tolerant of concentrated portfolios, with debt added on, when you invest in mature companies than you should be when investing in younger companies. By the same token, investors who find pricing mistakes in younger companies, but are daunted by the noisiness of their estimates or uncertainty about markets correcting, and thus are low on the conviction scale, can overcome their reluctance to act by spreading their bets across many such companies. As some of you may be aware, I did value SpaceX at the time of its IPO at about $100 a share, and as the price has drifted down towards that price, I may very well be faced with an underpriced stock (where the market price drops below $100). Given the uncertainty that is associated with my estimate, and you can see it in the simulation that I reported in my post, it is unlikely that I would ever have sufficient conviction to make SpaceX the biggest or only investment in my portfolio, but I stand ready to buy the stock as part of a portfolio, where it is one of many bets that I make on markets.

Lessons from Leo

I started this post with the story of Situational Awareness and Leo Aschenbrenner, but I spent most of it talking about investment conviction, what it is, its sources and consequences. I want to end the post by returning to Leo’s story and what we can learn from it as investors.

Lesson 1: Investment actions that are inconsistent with investment conviction risk ruin

I have no issues with Leo’s story of AI winning big in the near-term and buying the winners and selling the losers that will result. It is macro story investing, and it has worked for some in the past and failed for others, but if timed right, it can deliver significant returns. In fact, I will concede that Leo knows far more about AI than I ever will and is using that knowledge in constructing his AI story. I also have no bone to pick with investors using financial leverage to supercharge their returns, with low-risk investments, though I remain concerned that the (2 & 20) fee structure may lead them to use too much debt. My concern with Situational Awareness, as a fund, and this would have been true on June 19, even at it peak, is that combining a macro story about AI winning with maximal leverage creates a time bomb. The AI story, no matter how well told, has multiple obstacles to overcome, some related to business economics and some to politics and regulation, and is a risky bet, and it makes sense to fund it with significant amounts of debt. I know that there are defenders who will point to the fact that the fund, even after its markdown, was up substantially from its inception date, but the fact that leverage cut the fund’s life short only strengthens the case that if it had been run with less debt, it would have had a bad month in July, but lived to tell the tale and perhaps even deliver on its AI promise.

Lesson 2: Momentum is a wild card in every investment strategy, and you ignore it at your own peril.

It is a well-established finding in equity markets that momentum is one of the strongest forces moving markets and that it can overwhelm the best planned strategies of most investors. If you look at the composition of Situation Awareness portfolio through much of its rise and fall., the long positions were primarily in companies that benefit from the build-up of Ai architecture, selling their products and services to the hyper scalers and LLM companies, and the short positions were in software and other businesses that would be disrupted by the rise of AI. With both groups, Situational Awareness was taking bets that were in line with what the market was pricing in already, albeit in a more concentrated and leveraged form. While market observers were quick to link both the rise and fall of Situational Awareness to the AI story, you can make just as strong a case that much of that happened at the fund over its brief existence can be explained by momentum, with leverage acting as a super charger; continued momentum generated the outsized return though June 19, and the market reversal in July caused the correction.

Lesson 3: Humble money beats smart money

The legend of smart money persists in markets, where investors who are smarter than the rest of us, with access to information and capital that others do not possess, deliver supersize returns for themselves and those that they invite into their inner circle. That legend serves everyone’s interests, since the smart money uses its reputation to attract more capital and the not-so-smart money has someone else (hedge funds, insiders, activists) to blame for investment setbacks. While many money managers aspire to be part of the smart money group, most never make it into that rarefied circle, and those that do often have to pay their dues over long periods. Leo Aschenbrenner, in contrast, broke into the group in just a few months, perhaps helped by his pedigree as an AI insider and with an assist from his post on the coming AI revolution. The problem with smart money is that its self-regard makes its susceptible to attributing more precision to its own convictions, than merited by the circumstances, and that, in turn, results in over reach (portfolios that are much too concentrated or levered). Situational Awareness clearly overused leverage, and while some will attribute that to the youth and inexperience of its lead manager, it is worth remembering Long Term Capital Management, where John Merriweather, after a long and distinguished trading tenure at Salomon Brothers, aided by two Nobel Prize winners in economics, brought the fund to its knees by borrowing too much on risky trades. In a post from years ago, I drew a contrast between smart money and humble money, with the former including investors who attribute every basis point of excess return earned to their investing brilliance, and the latter open about the fact that their performance, no matter how stellar, has as much to do with being in the right place at the right time (being lucky) as it has to do with skill. Investors looking for someone to manage their money are likely to do much better with the latter than the former.

I hope that you don’t view this as a hit piece on Leo or AI, since that was not my intent. In fact, I hope that Leo persists and perhaps even comes back as a fund manager, since he strikes me as an original thinker who is willing to take a stand, both scarce qualities among active fund managers. I also hope that he has learned some lessons, especially on humility and restraint, for his next go around, and that he adopts a fee structure that gives his investors a chance of beating the market in the long term.

YouTube Video

Links to the Leo Aschenbrenner story

1. Situational Awareness: The Decade Ahead (Leo’s post on AI)

2. The 24-year old AI Wiz who counts Jane Street as an investor (The Wall Street Journal in early June, prior to blow-up)

3. Inside the Meltdown of a Wunderkind’s AI Hedge Fund (New York Times)

4. His Wedding Guests were arriving - Just as his $45 billion fund was falling apart (The Wall Street Journal)

5. What does the humbling of Leopold Aschenbrenner mean for the AI bubble? (The NewYorker)

The two stories may seem unconnected, but they have two common features. First, both actions (removing quarterly reporting requirements and reducing/eliminating Fed guidance), if carried through, will remove “news” that markets have become used to receiving and using to calibrate prices. Second, the arguments for and against each of these proposals have parallels. The advocates for removing quarterly reports argue that they feed into market myopia and increase short-termism in markets, and the supporters of less Fed guidance believe that this guidance creates gaming among traders and investors, increasing focus on the FOMC actions at the expense of fundamentals. The pushback against both proposals comes from those who believe that withholding quarterly reports and Fed guidance removes information that markets use to set prices, making these prices more volatile and less informative. As is almost always the case with these debates, there is both truth and hyperbole on both sides, and I will try to thread the needle.

Earnings Reports

Most investors and traders in equity markets, and especially so in the United States, have spent their investing lifetimes in an environment where companies not only release full financial statements every quarter, but do so with fan fare. As we will note in this section, that has not always been the case, even in the US, and has more recent origins, with setbacks, in many foreign markets.

The History of Earnings Reporting Periodicity

The Securities Exchange Act of 1934, which created the SEC, set up foundational annual reporting requirements (10-Ks) for publicly traded companies, modified in 1955 to require semi-annual reporting and in 1970, quarterly reports, within 45 days from the end of each quarter. That said, there have been forces that have induced firms to report earnings on a more frequent basis to investors well before these regulatory requirements were put in place. The first were the stock exchanges that imposed their own constraints, with the NYSE requiring most firms to report on a quarterly basis as early as 1939. The second was the recognition by firms that financial transparency (in the form of more frequent and more detailed financial reports) could make them more attractive to investors. As a consequence, it is estimated that in 1931, prior to either the SEC or NYSE mandating disclosure, more than 60% of publicly traded companies were already disclosing information on a quarterly basis.

The shift to more frequent reporting was slower in the rest of the world and has seen more reversals. Europe, for much of the last century, has a patchwork of rules, with some countries adopting stricter disclosure laws than others. The UK imposed mandatory quarterly earnings reporting in 2007, but allowed a shift back to semi-annual reporting in 2014, and the EU also followed a similar timeline, introducing quarterly reports in 2007 and withdrawing that requirement in 2013. In 2003, Singapore started requiring quarterly reporting for firms with market capitalization exceeding S$20 million, but in 2020, shifted away to requiring it only for a subset of firms with financial and regulatory concerns. Japan started its quarterly reporting requirements in 2003 as well, but it too reversed that requirement in 2024. In emerging markets, there are large variations across countries. In India, publicly traded companies are required to report their financials on a quarterly basis, and the same is true for many Brazilian and Chinese companies. In Africa, Nigeria requires quarterly reporting but South Africa has a semi-annual reporting mandate, though many companies voluntarily release quarterly financials; much of the rest of Africa has semi-annual reporting requirements.

In sum, the belief at the start of the twenty first century that the rest of the world would follow the US model of mandated quarterly reporting for publicly traded companies has not come to fruition, as many parts of the world have experimented with mandatory quarterly reporting, before abandoning it in favor of semi-annual reporting, for a variety of reasons. That said, it is worth noting that a significant percentage of firms voluntarily report their financial results on a quarterly basis, even when not mandated, albeit with different degrees of depth.

The Content of Earnings Reports

The debate about how frequently companies should report their financials misses a key detail related to what their financial reports include as content. Focusing on the US, for instance, the magnitude of quarterly earnings reports has increased over time, expanding from bare bones financial statements fin the 1970s to much larger documents that go well beyond financial statements today. In 1980, for instance, a typical quarterly earnings report contained 2000-5000 words, but by the turn of this century, those reports had tripled or quadrupled in size, and the trends continue. The graph below, for instance, looks at the growth in word count for the median quarterly and annual reports in the Russell 3000 companies between 2006 and 2020 (for quarterly) and 1994 to 2020 (for annual):

As you can see the number of words in both quarterly and annual reports has increased over time, and the bulking up of earnings reports can be explained by multiple factors:

• Accounting rule changes: Accounting rule writers have been busy adding more items to the list of required disclosures for public companies in the last few decades. Some of this increased disclosure (stock-based compensation, for example) reflects a changing business world and is merited, some is in reaction to a corporate scandal and if often knee-jerk and some, in my cynical vie, reflects accounting trying to be relevant to markets again.

• Macro events: In years of market crises, economic or political, you will see disclosures increase. In the graphs above, notice the spikes in 2008/2009 and 2020, the first in response to the 2008 banking crisis and the latter to COVID. Superimposing the effects of globalization, where a company finds itself exposed to problems in every corner of the world, it has added to the disclosure bloat.

• Legal Protection: One of the culprits responsible for disclosures bulk is the risk exposure section, where companies are required lay out an exhaustive (and exhausting) list of things that can go wrong in their business models. I have never found a risk disclosure useful in a valuation, as it seems to be written by lawyers with the objective of providing legal cover.

• Guidance: In the 1980s, quarterly earnings reports were focused on reporting on operations during the quarter in question and management was not expected to, and did not provide, guidance about future quarters. That started to change in the 1990s, especially with the passage of the 1995 Safe Harbor Law and Reg FD (which prevented companies from selectively leaking information to analysts), and surged through the second half of the decade, peaking in 2003, when more than 50% of all companies providing earnings guidance. Thankfully, the process has receded, with only a fifth of all firms now providing guidance with earnings reports, but it is undeniable that there is much more forward-looking components to earnings reports than used to be the case.

The bulking up of earnings reports is part of a phenomena that I term “disclosure diarrhea” and argue has undercut the usefulness of these reports, with more disclosure perversely making for less information.

The Earnings Game

To make sense of the arguments for and against quarterly earnings reporting, you have to get a measure of what happens leading into and out of these reports, in the “earnings game”. The process starts with analysts and investors making forecasts of what the earnings report will contain, almost always including estimates of the earnings per share, but often also containing estimates of expected revenues and even operating metrics (like margins) for high profile companies. The analyst forecasts, at least from sell side analysts, become quasi public information and are often aggregated and reported as consensus estimates by financial news services. Zacks, for instance, is one of the services that has been doing this for decades, but that information is now widely accessible on Google and Yahoo! Finance (with the estimates for Apple on July 27, 2026, for the September 2026 earnings report, shown below):

Yahoo! Finance for Apple earnings forecasts

These analyst forecasts, once made, are revisited, partly in response to company-specific news stories and partly to macroeconomic developments, and revised forecasts are provided, with services again tracking these revisions for trends (as you can see below for Apple, from Zack’s):

Zack’s Apple earnings revisions

As the earnings release date approaches, analysts continue to revise their estimates, and on the date of the announcement, the actual earnings per share is compared to the expected number, with higher (lower) than expected earnings labeled as positive (negative) surprises. The market price response is often consistent, with positive (negative) earnings surprises translating into increases (decreases) in stock price. The graph below, while dated, looks at stock price responses to earnings surprised classified into ten deciles (from most positive to most negative):

There is some evidence that the market responses to earnings reports have become more muted over time, perhaps because of public access to analyst forecasts and revisions. The link between earnings surprises and stock price changes has become the reason why analysts spend as much time as they do, forecasting earnings per share in the next quarterly report, and why traders focusing on the same metric. There is another aspect of the market reaction to earnings surprises that becomes grist for the trading mill, and it comes from the price drifts in the days after earnings are released, with positive (negative) surprises followed by upward (downward) drifts. While the price drift is small, it may still be large enough to make a difference in active trading, where winning by inches is still winning.

As with almost everything else that is market-related, there are no easy wins in this game, and as more and more people play the earnings forecasting game, new wrinkles have emerged. First, companies have learned to use the flexibility embedded in accounting rules to find ways to beat analyst estimates, with tech companies, in particular, standing out. That earnings gaming plays out as a disproportionately large number of positive earnings surprises (at least among the S&P 500 companies, broken down by sector), as is clear from earnings surprises at the S&P 500 companies in the second quarter of 2026:

Source: Factset

Second, as companies routinely beat analyst estimates, markets readjust, creating the phenomenon of whispered earnings, where investors build in the expectation that a company that has historically delivered earnings that are 5% or 10% above estimates will continue to do so, and a lesser number is a negative surprise. In the graph below, I look at the market price reaction to earnings surprises in the second quarter of 2026:

Source: Factset

As you can see, the linkage between earnings surprises and price reaction is weak, with a significant subset of positive surprises resulting in price drops.

The Bottom Line

Much of the debate about whether the US should shift away from quarterly to semi-annual reports can be boiled down to what you think about the time and energy investors and companies spend playing the earnings game, and where that time and energy will be spent in the absence of quarterly reports. Those who are advocates for less frequent reporting are of the view that the earnings game, focused as it is on next quarter’s earnings estimates and whether the company can beat them, contributes to short-termism and distracts from fundamentals. Those who are pushing for preserving the status quo (of quarterly reporting) believe that removing quarterly reports will just shift the game, perhaps more intensively, into the semi-annual reports and that there is value to long term investors from having quarterly reports, gaming notwithstanding.

There is another issue that comes up in the context of quarterly reporting, and what would happen if these reports did not exist. Legal strictures notwithstanding, insiders (from within and outside the firm) trade and make money on material information that they have access to, but the public does not. Removing quarterly reporting will create more of an opening for insiders to make money at the expense of public market investors, and while inside trading may contribute to making prices more informative, it also adds to the sense that financial markets are an unfair game.

I am an investor, and I think that there is a compromise solution that draws on both sides of this argument. I like quarterly reporting for two reasons.

1. There is information in those reports that allows me to update my company valuations, though for many companies, the marginal impact of a quarterly report on value is small.

2. While I have no interest in playing the earnings game, the price corrections that happen around earnings reports serve two purposes. For companies that I have a position in, they can operate as catalysts, bringing down (up) the stock price of over valued (under valued) companies. At the same time, almost all of the information that I find useful in an earnings reports is in the financial statements and footnotes, not in the lengthy discussions of risk exposure or in the management guidance, and I would welcome an elimination of these sections and a slimming down of these reports.

Note that none of my arguments for preserving quarterly reporting are about short-termism, and that is intentional. First, I am not sure what short-termism even means, since the cynical answer seems to be that any market movement away from your preferred price direction is short term, and any movement in your favor is indicative of market wisdom. Second, I believe that most market participants, and this is true across time and markets, trade to make money in the near term, and that there is nothing that regulators or rule writers can do to alter this dynamic. In fact, the magic of markets is that millions of trades motivated by opportunism and the short term can still yield a price that is long term and rational. Finally, it remains true that if we were all long-term investors who traded only when the fundamentals drove us to do so, markets would be less liquid and transactions costs would increase; short term traders provide a market service and supply liquidity that we all (including long term investors) benefit from.

I hope that the SEC preserves the current quarterly reporting requirement, while scaling back the volume of disclosure, but if it decides otherwise, it will not materially change much of what I do. I will miss the quarterly updates more with younger, higher-growth firms, where the operating metrics (revenues, margins etc.) can change quickly over short periods, but it is my guess that many of these firms will voluntarily continue the quarterly reporting tradition.

Fed Guidance on Rates

For most investors who started investing after 2008, the Fed, in particular, and central banks, in general, have loomed large in the investing process. Many investors attribute the low interest rates after 2008 almost entirely to Fed actions, and by extension, blame the Fed for the higher rates since 2022. I have long argued that not only is this perception incorrect, but that it is unhealthy for investors to view the Fed as either savior or villain.

A Short (and Personal) History of the Fed

I started in equity markets in the 1980s, when Paul Volcker as the chair of the Fed played a central role in getting inflation back into check. I might have been ignorant, but I did not know the names of any of the members of the Federal Open Market Committee and had no idea when they met. Changes in the fed funds rate, the only rate effectively controlled by the FOMC, would percolate their way into markets, but I don’t remember them being central to equity market movements.

Volcker was followed by Alan Greenspan, and while he acquired rockstar status (at least among investors) in the late 1990s, his views on rates were superseded by his views on equity investors (and their irrational exuberance). The FOMC met eight times per year during that period, and you can access the meeting minutes and actions on the Fed website here, but it stayed away from explicit guidance about future rate changes, choosing to send subtle hints instead.

The sea change in Fed behavior and centrality occurred with the 2008 market crisis, when the Fed first introduced explicit guidance noting that rates would stay low “for some time”, and it has largely continued that practice through the stewardships of Bernanke (2006-2014), Yellen (2014-2020) and Powell (2020-2026). Along the way, its place in markets has changed, as both bond and equity investors have become focused on the Fed as the arbiter of interest rates and director of the economy.

The Fed’s Powers (and Powerlessness)

To understand the extent and limits of the Fed’s capacity to guide rates and the economy, it is useful to begin with an understanding of what it does. Through its twelve districts that span the United States, the Fed collects information on almost every aspect of the economy, from price pressures building on consumers and producers to the pace of economic growth. While there are other government agencies that also track these statistics, it is undeniable that the Fed has a big picture view and access to more data than any other government agency. The Federal Open Market Committee, composed of all of the members of the board of governors and representatives of the district presidents, sets Fed policy on open market operations (where the Fed buys and sells US government securities), the size of the Fed’s balance sheet and the Fed Funds rate (an overnight rate at which banks can borrow and lend their reserves). In addition to the FOMC providing policy direction on inflation and the economy, the Fed chair testifies to Congress every six months, facing and answering questions from legislators.

As the key interest rate set by the Fed, the Fed Funds rate often acquires an outsized role and there are good reasons to pay attention to it. First, it operates as a signal of what the Fed is seeing in the data it has collected on the economy, with an increase (decrease) in rates indicating that it sees higher (lower) inflation and an overheated (slowing) economy. Second, there are interest rates that are directly tied to the Fed Funds rate, where changes percolate down to businesses and customers; the prime rate and some credit card and CD rates move with the Fed Funds rate. That said, I believe that Fed’s capacity to affect interest rates is far more limited than most believe, for two reasons. First, while there is positive correlation between Fed Funds rates and short-term market-set rates (like the US treasury bill rate), there is as much evidence (if not more) that the latter lead the former, rather than the other way around. Put simply, Fed funds rates tend to be increased (cut) after short term treasury rates have gone up (down), suggesting that the Fed is mimicking the market. Second, the relationship between Fed Funds rates and long-term market-set rates, which drive asset valuation and affect borrowers more, is even weaker. To back these contentions, I chart the effective fund funds rate, the three-month US treasury bill rate and the 10-year treasury note rate on a monthly basis from January 1962 to June 2026:

Download data

At the bottom of the graph, I have a table where I look at the data on a quarterly basis, and break it down into three groups - quarters where the fed funds rate decreased, quarters where it increased and quarters where it stayed unchanged. With both fed funds rate increases and decreases, you can see that the link with short term rates is stronger, and with both short term and long term rates, the bulk of the change in rates happens prior to or in the quarter that the Fed Funds rate changed, but there is only a mild spill over into the quarter after, with three month rates, and almost no spillover, with long term rates. Put simply, baed on this history, it looks like changes in fed funds rate are less signals of future movements in interest rates and more reflectors of changes that have already happened.

The Fed’s weaknesses in setting interest rates also plays out in its capacity to alter the trajectory of the real economy. While there are clearly periods that you can point to where Fed actions have had a material impact on he economy, with the Fed Fund rate was hiked to 20% under Paul Volcker in 1981, and triggering a deep recession, being a prime example, the link between Fed Fund rates and economic growth remains tenuous. In the graph below, I look at the changes in Fed Funds rates and real GDP growth in the quarter leading into, the quarter of and the quarter after the change:

Download data (FRED)

Again, there is little backing for conventional wisdom, which is that fed tightening (by raising the Fed funds rate) leads to drops in real growth (or even recessions) and that fed loosening (by lowering the Fed funds rate) is a signal of higher economic growth in the future. In fact, the more general conclusion that one can draw from the data is that the fed effect on the real economy has been more “meh” than “wow”.

The gap between investor perception on what the Fed can control on interest rates and the economy and its actual powers is not just wide, but potentially dangerous. From a policy perspective, it can lead to perverse actions, where central banks are pressured to lower the rates they control (like the Fed Funds rate) in the face of high inflation, leading to even higher inflation in the future. From an investor and business perspective, the focus on what the Fed is doing or will do can take attention away from the fundamentals, especially inflation, that ultimately drive both interest rates and growth.

Download data

As you can see, much of the variation in long term interest rates (with the ten-year US treasury rate standing in as proxy) can be explained by movements in inflation and real economic growth over time, not Fed action or inaction.

The Warsh Doctrine?

All Fed chairs have had to wrestle with the problem of being perceived as all-powerful, when their true powers are limited, but Kevin Warsh is perhaps more exposed than any of his predecessors. The market fixation with the Fed is now deeply embedded in market, and there are politicians on both sides of the aisle who seem to think that Warsh can bring rates (mortgage, treasury) down to 2% or lower, if he so desires, when the truth is that with inflation expectations running at 2.5-3%, there is no chance of that happening.

The pathway out of this problem will be long and there will be pushback, but the end game should be a world where you see and hear from the Fed less, not more. I do believe that the decision to reduce or withhold guidance is a good first step, and it has to be followed by more open humility from the Fed (and from Warsh) about the limits of its powers and honesty about how frequently it follows markets, rather than leads them. In the context of today’s (July 29, 2026) decision by the FOMC to leave rates unchanged, for instance, the subtext is that while inflation is running hotter than desired (3% or more, as opposed to the targeted number of 2%), much of that inflation is being driven by a war and its effect on oil prices

There will be some who feel that markets will be lost without Fed guidance, but I don’t think so. After all, financial markets set interest rates and stock prices before the guidance era, and did a pretty good job. In fact, I think that the surge in guidance from the Fed has led many in markets to abdicate their responsibility for paying heed to fundamentals and gauging what interest rates should be.

Conclusion

If there is a takeaway from this post, it should be that there is nothing inherently good about having more disclosure. In fact, there is a tipping point, where information overload can cause investors to behave in perverse ways. Thus, I am less of an absolutist about the quarterly versus semi-annual reporting debate than some, though my view is that rather than reduce the frequency of reporting, the SEC should be looking at slimming down reports, by replacing one-size-fits-all disclosure requirements with targeted disclosures and keeping the focus on reporting what has happened rather than prognosticate about the future.

With the Federal Reserve too, I think less is more is a better strategy - less guidance from the Fed about what it will do in the future, less opining from FOMC members about interest rates and the economy and less attention to FOMC meetings and the smoke signals that emerge from these meetings. Markets will step in to fill the vacuum, and that is good not just for investors but for the Fed, since its decisions are informed by those market judgments.

YouTube Video

Data

1. Fed Funds Rates and US Treasury rates, by month (1962-2026)

2. Fed Funds Rates and Real GDP growth, by quarter (1962-2026)

3. Intrinsic risk free rates and 10-year Treasuries (1954-2025)

Country Risk: Relevance

In my years as a business school student, country risk was given short shrift and I don’t remember spending much time talking or thinking about it. Part of the reason was that business school education was dollar-centric and built on the presumption that most graduates would go to work in New York, London or Tokyo, and have little need to confront country risk on a day-to-day basis. For those who raised country risk as an issue, the response was that you could, as a company or investor with global exposure, diversify it away. Both presumptions were wrong even then, and have become even more flawed over time as we have sold both companies and investors on the benefits of globalization.

For businesses, the exposure to country risk comes from both the revenue side, as larger portions of every company’s revenues come from foreign markets, and the cost side, as production gets outsourced to locales overseas. That exposure tends to increase as companies scale up, and is higher in some sectors than others; technology companies, for instance, get far more of their revenues from other non-domestic markets than manufacturing or service businesses. Outside of utilities (power, water), it is rare for a company to be entirely domestic-focused on both its revenue and cost sides. For investors, the initial draw of investing in foreign markets might have been diversification but the greater pull has come from greed, i.e., the belief that you can higher returns in the rest of the world. That process was accelerated by the creation of investment vehicles (index and mutual funds) that made investing overseas easier, the lowering of transactions costs across markets and a greater standardization of financial statements and disclosure across the globe. The home bias in portfolios, i.e., the skewing of portfolios towards domestic market investments, has not disappeared but it is lower than it was at the turn of the last century.

The notion that country risk is diversifiable, i.e., that if you are operating or investing across the world, the risks will average out across countries, has been undercut by the increased correlation across global equity markets, and especially so during market crises (which is when you care the most). At the risk of being hyperbolic, there is no place to hide from country risk, for either businesses or investors, and ignoring or dismissing country risk is not an option. I discovered this truth in the 1990s, when I found myself in need of a mechanism to incorporate country risk into my corporate financial analysis and valuations, and the process that you see described in this post was born from that need. I would hasten to add that the process that I describe has very little intellectual firepower behind it, puts pragmatism ahead of theory and most importantly is a work-in-process.

Country Risk: Drivers

I don’t think that there would be much disagreement, if I assert that it is riskier to invest in some parts of the world than others, but there is likely to be plenty of disagreement on why there are risk differences and which parts of the world are riskiest. In the broadest sense, I argue that variation in business risk across countries can be traced to four factors - the political structure of the country (democracy vs authoritarian), the prevalence of corruption in the country (operating as a hidden tax and distorting business outcomes), the extent of violence in the country (from internal and external forces) and the strength of the legal system in enforcing property rights and contractual obligations.

On the political risk front, I looked at the EIU’s Democracy Index, a composite score measuring both political freedom and protections of civil liberties, with the caveat that any index that tries to measure these will make subjective judgements that not everyone will agree with. In their most recent update, here is what the EIU scores looked like around the world:

Source: Economist
Low (High) score: Least (Most) freedom

Based on these scores, the tilt towards authoritarianism has increased over the last decade, with only 7.3% of the world’s population living in democracies at the end of 2025. Note, though, that there is still an open question of whether businesses and economies do better under democratic than authoritarian regimes, and the answer in the research is at best a “maybe”. From a risk perspective, democratic regimes create more continuous risk for businesses, with elections bringing regulatory and rule changes to economies, than authoritarian regimes, where governments can promise more continuity in policy, but when change does come to the latter, it is more likely to be large and wrenching.

Corruption is a fact of life in much of the world, and businesses often have no choice but to pay the price to survive and grow. Transparency International, a global coalition against corruption, tries to capture the extent of corruption, comes up with corruption scores for countries, with lower scores indicating less corruption, and the most recent edition contains the following:

Source: Transparency International
Low (High) score: Most (Least) corruption

Northern Europe has the lowest corruption scores, followed by Canada, United States and Australia, but large portions of Africa have high exposure to corruption, with Latin America and much of Asia falling in the middle.

Living in the midst of violence takes a toll, and that toll is extracted from businesses that try to operate in its presence. Vision of Humanity computes peace scores for countries, measuring exposure to both violence within the country as well as from wars and terrorism. The most recent peace scores are reported below:

Source: Vision of Humanity
Low (High) score: Most (Least) peaceful

Canada, Australia, Japan and much of Europe score high on the peace dimension, and while Latin America and Africa score lower, there are portions of each continent that are more peaceful. The Russia-Ukraine war has created a huge area of violence across Eastern Europe and Russia, and exposure to gun violence creates a drag on the United States.

Businesses are dependent on the legal system to enforce property rights as well as contractual obligations. Countries that have legal systems that are either capricious on these fronts, or hopelessly slow in acting, create challenges for businesses that operate in them, creating both costs and risks that they otherwise would not face. Property Rights Alliance is an entity that tracks international property rights across the world, and in their most recent update, their property rights scores by country are captured below:

Source: International Property Rights
Low (High) score: Least (Most) property rights

There are wide differences across regions, when it comes to legal and property rights, with Latin America, Africa and Asia lagging and Europe, Australia and much of North America leading.

There is one final dimension that I have added to country risk in recent years that captures exposure to climate risk. While there are many different entities that measure this exposure, each one with its own skews, the map below which shows the climate risk exposure, by country, from GermanWatch:

Source: GermanWatch
Low (High) score: Least (Most) affected

There are two reasons why climate risk has not become a bigger topic in country risk discussions. The first is that there is no part of the globe that is unaffected, making it less of a differentiator across countries on the risk dimension. The second is that climate risk, by itself, is an abstraction for businesses, until it starts affecting the bottom line, and while there are individual companies that are being impacted, the aggregate effects, at least at the moment, are not big enough to change the discussion.

Country Default Risk

While country risk is determined by multiple factors, the challenge that businesses is in consolidating all of those risks into one number. The market that does this most directly is the debt market, where, when countries (sovereigns) seek to borrow money, lenders determine the interest rates to charge them, based upon perceived default risk. To understand why lenders worry about default with sovereign debt, you can start by looking at the history of sovereign defaults in the graph below:

Source: BoC & BoE Sovereign Default Database

Debt defaults, which soared in the 1980s and 1990s, have been lower in this century, with a shift away from loan defaults (where banks are usually the lenders) to defaults in the bond market. It is also worth noting that a non-trivial portion of sovereign defaults in each year are local currency defaults, indicating that for some borrowers, the costs of defaulting are viewed as smaller than the costs of inflation arising from printing more currency to pay off debt. Over time, Latin America has been the epicenter for sovereign default, but at the end of 2023, sovereign debt in default had a wide geographical spread:

Source: BoC & BoE Sovereign Default Database

The most widely accessible measures of sovereign default risk remain sovereign ratings, with ratings agencies operating as (imperfect) arbiters. At the start of July 2026, the graph below reports the sovereign ratings for all rated countries, from S&P, Moody’s and Fitch:

Source: Multiple public sources

As you can see, the ratings agencies mostly agree on their assessments of default risk, and sovereign ratings are correlated with the risk drivers (politics, corruption, violence, legal system) that we outlined in the last section. I do believe that ratings agencies, notwithstanding the critiques of bias and mis-measurement leveled against them, do a reasonably good job in their ratings assessments, but they are often slow to act, when confronted with change.

The sovereign CDS market offers a market-based alternative for measuring sovereign default risk, with investors making assessments of how much they would demand to insure against sovereign default in the form of (annualized) spreads. In the graph below, I list 10-year sovereign CDS spreads as of July 1, 2026:

Source: Bloomberg

Note that sovereign CDS spreads are available for only 84 countries, and that there are swaths of the world (Central and North Africa, frontier markets) where they are not available.

Country Composite Risk

When you lend money to governments or buy government bonds, sovereign default risk is your key concern, and both sovereign ratings and CDS spreads try to measure that risk. When running a business in a country, you are exposed to a much wider range of risks, and measuring exposure to those risks may require different measures. One alternative is country risk scores, where services evaluate how countries measure up on different risk drivers, and come up with composite scores for these countries. In the table below, I report the country risk scores from two services - Political Risk Services (PRS) and the Economist (EIU), at the start of July 2026:

Sources: EIU (Economist) and PRS

The table illustrates three problems that you face with political risk scores. The first is that the scoring is idiosyncratic, with the Economist going from low scores for the safest countries to high scores for the riskiest, and PRS doing the reverse. The second is that each service picks different factors to consider, and different weightings, leading to scoring divergences that sometimes confound; PRS, for instance, ranks the United States as riskier than Ghana, on a composite risk basis. The third is that the scores, by themselves, are difficult to convert into inputs in financial analysis, either in cash flow or discount rate adjustment.

It is to combat the third problem that I started estimating country equity risk premiums, and while the details of the process and the data that I use have changed over the last three decades, the basic structure has remained unchanged. I start with an estimate of the equity risk premium for a mature market, and build a country risk premium, if needed, for riskier countries. Until 2025, I estimated the mature market premium by computing an implied equity risk premium for the S&P 500, and using that as the base, arguing that the US, as a Aaa rated country (at least according to Moody’s), represented a mature market. The Moody’s downgrade for the US, from Aaa to Aa1, has thrown a wrench into that approach, requiring adaptation. In response, I now start with an estimate of the implied ERP for the S&P 500, but then adjust that estimate for the default spread (based on the Aa1 rating) for the US, with the resulting values at the start of July 2026 below:

Spreadsheet: https://pages.stern.nyu.edu/~adamodar/pc/implprem/ERPJuly26.xlsx

As you can see, with the S&P 500 at 7499.36 on July 1, 2026, the implied equity risk premium for the United States is 4.42%, and netting out the default spread of 0.22% for the Aa1 rating yields a mature market premium of 4.20%.

To estimate country risk premiums, I start with the sovereign ratings for rated countries and convert those ratings into default spreads. To adjust for the higher risk associated with equities, relative to government bonds, I estimate a composite measure of that relative risk, by scaling the volatility in an emerging market equity index to the volatility in a emerging market government bond ETF, and scale the default risk up with this relative risk measure (1.55 in July 2026) to get country risk premiums:

For the two dozen countries that have no sovereign ratings, I adopt an even more makeshift approach, where I used political risk scores for these countries, and then looked for rated countries with similar scores. The table below has equity and country risk premiums, by country, for all of the countries that I evaluated in July 2026:

Download data

I did post an earlier version of this table a couple of weeks ago, but the numbers that I reported reflected in incomplete update of sovereign default spreads, and this table (and the data on my webpage) now reflect the corrected (and lower) spreads. (As a solo act, I am deeply grateful for the checking that those who use my data do, and thankful when they point out mistakes that I have made.)

Company Exposure to Country Risk

If you buy into my argument that every company has a narrative, and it is the narrative that drives its value, it is worth considering where country risk fits into that narrative. The answer, I believe, comes from looking at where the country in question falls in the life cycle:

The message from this life cycle view is a sobering one, especially for those analyzing companies that operate in very risky countries, since the narratives for these companies implicitly or explicitly incorporate a country risk component. You cannot value a Venezuelan company without taking a strong view about Venezuela, or even an Indian and Brazilian company without an India or Brazil country story underpinning value. In contrast, you may be able to value US and European companies, without explicitly considering the evolution of country risk in those parts of the world.

When looking at an individual company, I believe that country risk exposure comes less from where the company is incorporated and more from where it operates. It is undeniable that companies around the world have substantial exposure outside their domestic markets, and that exposure has increased over time. In the graph below, I look at the revenue breakdown of companies in four indices - the S&P 500 (US), the FTSE 100 (UK), the Nikkei 225 (Japan) and the Sensex (India):

In every single index, companies that comprise that index get a significant portion of their revenues from outside the domestic market. Looking across sectors, exposure to foreign markets varies widely with technology companies often generating more than half of their revenues outside their domestic settings. I believe that equity risk premiums for companies should reflect exposure to foreign markets, though it is worth debating how best to weight that exposure - revenues work well for consumer product and service companies, production works better for natural resource companies and a mix of revenues and production may be the right choice for manufacturing companies:

With this framework, you can see why almost all analysts will confront country risk, sooner or later, no matter where they operate in the world and which companies they analyze.

For companies, country risk will also come into play when faced with capital budgeting decisions, where they need estimates of hurdle rates for individual projects, to decide where to invest. For a multinational operating in many businesses, the project cost of equity will have to then also reflect the business the project is in, in addition to country risk. Thus, the cost of equity for a Siemens Appliances for a project in India should reflect the beta for the appliance business, in addition to the country risk for India. In contrast, a Siemens power tool project in Hungary should be computed using the beta for an power tools project and the country risk for Hungary. It is also possible that country risk is not easy to isolate, if the production facilities are in one country but revenues are generated in another. If the Siemens appliance factory in India will be producing products that will be sold in Japan, should we be showing the country risk of India or Japan in the cost of equity calculation? The answer, as was the case in the earlier section on valuation, is that it depends on where the company sees risk coming from. If the risk is that production will be delayed or disrupted by political and economic risk in India, it is Indian country risk that should be looked at, but if the primary concern is that revenues in Japan will be volatile because of economic conditions there, it is Japanese country risk that matters more. If both risks are considerations, you should use a weighted average of Indian and Japanese country risk.

Currency Questions

For some of you, it may seem odd that I have spent almost an entire post talking about country risk without bringing up currencies. The reason is simple. Currencies are measurement mechanisms, and while they may be affected by the same political and economic factors that drive country risk, they don’t determine country risk and in my view, should not command risk premiums, on their own.

It is true that hurdle rates are affected by both the equity risk premiums that you estimate and the riskfree rate that you use, and that riskfree rates vary across currencies. In the figure below, I estimate riskfree rates in about 40 currencies, where a local-currency government bond rate is present, and I adjust that government bond rate for the default risk of the government in question:

When estimating the cost of equity for a Turkish project or company in Turkish lira, we start with a riskfree rate in excess of 20% and build on it, by adding equity risk premiums to it, but the cost of equity for the same project or company in Euros will begin with a riskfree rate close to 3% (the German Euro bond rate) and arrive at a much lower number. While this may sound farfetched, the value that you derive for the project or company should be the same using either currency, if you are consistent about estimating your cash flows in the same currency:

Since much or almost all of the differences in riskfree rates come from inflation differentials, matching the high Turkish lira discount rate with a high growth in cashflows in Turkish lira, and the low Euro discount rate with the low growth in cashflows estimated in Euros will yield results that are consistent.

If you do want to estimate riskfree rates in currencies where there is either no local currency government bond that is traded or where you mistrust the government bond rate, because of light trading or government intervention, the fact that riskfree rate differences across currencies can be tied to differential inflation can be used for estimation; the riskfree rate in any currency can be computed from a base currency (dollar or Euro) riskfree rate and the difference in expected inflation between the local and base currencies:

Put simply, if the expected inflation rate and riskfree rate in US dollars are 2.5% and 4% respectively, and the expected inflation rate in Brazil is 10.5%, the riskfree rate in Brazilian reais should be roughly 12%. The implication of this approach is that currency pegs, when they do exist, will hold only if the inflation in the pegged currency matches or is close to the inflation in the index currency to which it is pegged. It is true that the estimates of riskfree rates will only be as good as the expected inflation rates that are embedded in the estimation, but the good news is that being wrong on expected inflation will be largely offsetting, since both your cashflows and your discount rates will be wrong in the same direction; if you underestimate expected inflation, you will underestimate (overestimate) your riskfree and hurdle rates, but you will also underestimate (overestimate) your expected growth rate in cash flows.

Conclusion

One of the side effects of the rise of globalization is that there are fewer and fewer companies that are entirely local-country focused in both their revenues and production, and as a result, almost every business and investor is exposed to risk in other parts of the world. The problem with measuring country risk is that while its consequences are economic, it has its sources in history, politics and governance structures. The measures of country risk, whether they be entity-based like sovereign ratings, or market estimates like sovereign CDS spreads, reflect this interplay.

I confess that I have made simplistic assumptions and cut corners in my attempt to estimate equity risk premiums, by country, and there will be individual countries, perhaps even your own, where you might disagree with my assessments. As I noted earlier, my estimation approach remains a work-in-progress and I am always open to suggestions on how to estimate these premiums better, but keep in mind that whatever those improvements may be, they will have to work across 180 countries.

YouTube Video

Papers on country risk and equity risk premiums

1. Country Risk Premiums - The 2026 Edition (July 2026)

2. Equity Risk Premiums - The 2026 Edition (March 2026)

Data

1. Equity Risk Premiums, by country - July 2026

2. Implied Equity Risk Premiums for the S&P 500 (Annual since 1960 & Monthly since Sept 08)

3. Inflation-based riskfree rates, by currency - July 2026

Spreadsheet

1. Implied ERP for S&P 500 (July 2026)

I am skeptical, since I see a combination of hidden agendas and misguided views about investing behind each of the three groups that are most vehemently against inclusions. First, you have a cadre of active investors, many of whom have been left bruised by a losing battle that they have waged over the last two decades against passive investments (index funds and ETFs), who view inclusion in the index as a fait accompli, and present this as an added risk to passive investing that can be avoided by paying these professional money managers to avoid that risk. Second, you have investing experts and academics who claim to be looking out for for retail investors and retirees, and view including these big, money-losing companies in indices as dangerous for these small investors, partly because they may not be aware of their exposure and partly because they should not be investing in these types of companies. Third, you have politicians, normally not founts of investment wisdom, speaking out about how including these large companies in government pension funds will reward billionaires, who are the villains in their storylines. In this post, I will try to step back from the heat and try to cast some light on the question of index inclusion, starting with an understanding of how indices are constructed before moving on to the roles they perform in markets and ending with a discussion of whether and how inclusion of these companies will affect the passive versus active investing debate.

Index Construction - Inclusion, Weights and Returns

Indices have been around almost as long as assets have been bought and sold in markets, but it is undeniable that the extraordinary growth of financial markets in the last few decades, across geographies and asset classes, has added rocket fuel both to the number of indices in existence as well as their visibility. But what is it that sets one index apart from another, and why do indices that purport to measure the same market sometimes move in different ways? To understand the answer to these questions, we need to deconstruct indices and see how they are put together:

• Constituents: The first and perhaps most critical determinant of an index are its constituents, and what determines their inclusion. Take, for instance, the S&P 500, which Standard & Poor’s (its creator) describes as the “gauge of large-cap US equities”, and is without doubt the most widely tracked and followed index in global markets. As its name indicates, this index has five hundred of the largest market-cap companies listed and traded in the United States, with caveats on inclusion relating to listing age (listed at least a year), liquidity (measured by looking at shares that are available for investors to trade in the market, i.e., free float) and profitability (positive profits in the four quarters leading into the listing). There are local indices that exchanges (NYSE, NASDAQ), equity markets in other geographies (the Bovespa for Brazilian stocks and the Sensex for Indian stocks) and individual sectors or industries. Across asset classes, there are indices for fixed income, as well as for real estate, fine art and crypto currencies.

• Weights: You can have two indices that contain the same companies that register very different results over time, depending on how these companies are weighted, with three common choices. The first is to weight every company in an index equally, with the benefit being simplicity, but the cost being that to the extent that companies in an index have very different sizes, the performance on an equally weighted index will not capture aggregate market performance, because it will be skewed towards smaller companies. The second, and one used by some older indices like the Dow 30, is price-weighting, where the companies with the highest price per share are weighted more than companies that have lower priced shares. I cannot think of a single redeeming quality to price weighting, since it measures very little of consequence, and suffers from breakdowns, right after stock splits. The third and most widely used mechanism for construction indices is market capitalization, with tweaks sometimes added on for float (traded shares). The S&P 500, as I noted earlier, uses market capitalization, based on free float, to weight companies and as a consequence, Meta punches in below its true weight, since the bulk of class B shares (which are voting shares held by Zuckerberg) are not counted, as does Walmart, where some family-controlled holdings are treated as non-traded.

• Index level mechanics: Once constructed, an index has to be measured, and to the extent that these indices are designed to capture market prices, the first step is creating a mechanism for converting market prices on the constituents to an index level. Consider, for instance, the S&P 500 which ended trading on June 15, 2026, at 7554.29, and relating that number to the market capitalization of the companies that make up the index. At close of trading on June 15, 2026, the cumulative float-adjusted market capitalization of the 500 companies in the index was $63,498.44 billion and the index units for the conversion can be computed as follows:

Index units = Index level / Float-adjusted market capitalization = 7554.29/ 63498.44 = 0.1190

Note that there is no intuitive significance to the index units standing alone, but its movements over time can be an indicator of changes happening at companies, because of issuances and stock buybacks, as well as changes in index constituents. If asked to compute earnings or dividends on the S&P 500, these index units come into play again, when converting the aggregated dividends and earnings across all of the S&P 500 companies into index dividends and earnings. In 2025, for instance, the aggregated dollar dividends on the S&P 500 was $664.90 billion, and multiplying that value by the index units (0.1190) yields an index dividend of 79.12 for the year.

• Price updating: While index levels are starting points, most investors track indices for changes in the index, with increasing stock prices translating into higher index values. Indices that track publicly traded stocks, like the S&P 500 and the Dow 30, should adjust instantaneously as the prices of their constituent companies change during the course of a trading day, making the index a real-time measure of market movements. Indices that capture only price changes miss the other component of returns on a stock, which is dividends, and constructing an index that incorporates dividends paid on a continuous basis does take work and requires assumptions about whether the dividends are reinvested in the index or extracted by investors. Though not as widely disseminated as the pure-price version, there is a variant of the S&P 500 that computes the total return on the index, with dividends included. Indices of assets that are not continuously traded, most notably real estate (like the S&P Case-Shiller home price index), try to overcome the absence of price data on the assets by extrapolating from the pricing of the subset of assets that get traded, leading to noisier estimates for index value and lags in price adjustment.

• Index changes (inclusions and exclusions): Even the best constructed indices have to confront change and have mechanisms to deal with that change that are transparent and quick to put into practice. Some of that change will come from companies being removed from public markets, either because they are acquired, taken private or because of bankruptcy. Some change will be caused by new companies being listed on the market or some will be created by changes in market cap in companies that bring them into contention for inclusion in the index, either because the market cap has risen (making them large enough to qualify for a large cap index) or has dropped, removing them from large cap status. Since you do not want abrupt changes in the index level coming just from replacing a company with a low market cap with one with a much higher market cap, the adjustment has to come from changing the index units. Thus, assume that a company goes public with a trillion dollar market cap and that it will be replacing a company with a one-billion market cap, the adjusted index units for the S&P 500 will be as follows:

  Index units = Index level / Float-adjusted market capitalization + Market cap of added firm - Market cap of eliminated firm = 7554.29/ (63498.44+ 1000 -1) = 0.1171

  This will then percolate through into the index earnings and dividends estimates, for the index. Note that while the index level will be unchanged by the addition of the trillion dollar company, the other components that it brings with it, including higher growth and perhaps negative earnings, will alter the fundamentals of the index going forward.

With these index mechanics in mind, it is quite clear that if S&P does include SpaceX, OpenAi and Anthropic in the S&P 500 index, the index will not change at the time of the replacement, but it will change the index fundamentally going forward, bringing in more risk, a near term hit to earnings and perhaps a long term increase in growth.

The Index End Game

When indices were first created for markets, their primary purpose was to create composite measures of market performance, with a single number (the index value) capturing the performance of a much larger group of assets. Over time, though, the use of indices has expanded, first as proxies to assess the performance of active investors, to see whether they over or underperformed, and then as investment vehicles, with the advent and growth of index funds and ETFs.

Measurement

I started in equity markets in 1981, and at that time, indices were primarily measures of market performance. At the time, for most investors without intraday access to markets and without financial news channels, news of market performance, on most days, was a snippet on the evening news, where the anchor would mention the market’s change during the day, usually using the Dow 30 as a stand in for the market. Indices continue to perform the measurement role, though our access to data has changed dramatically, with real time updates on our devices occurring all through the day. From the measurement perspective, it is worth looking back at index construction and looking for indices that best capture what you are trying to measure. The reason that the S&P 500 has acquired primacy is that while it includes only 500 companies in a US equity market that has almost 6000 publicly traded companies, the fact that these are the companies with the largest market capitalization means that the index represents more than 80% of the market capitalization of all US equities, and as a result, it is the single best proxy for aggregate equity market performance in the index. That said, a different framing of the measurement question can lead you to a different index choice. Thus, if you are trying to measure how the average US equity did during a period, you may be better served using an equal-weighted equity index for that measurement.

Performance Evaluation

Investors who trust professionals to manage their money, and pay them for their services, either as up-front entry fees or in annual management expenses, are entitled to wonder whether they are receiving a compensatory benefit, in the form of higher returns. It should not be surprising that comparing a mutual fund’s returns to the returns on an index becomes a proxy for fund performance, and in the early years of performance evaluation, the S&P 500 became the default comparison index. Used in that context, one of the most jarring numbers in active investing is the percentage of active large cap funds that earn returns that are lower than the S&P 500, each year for the last two decades:

S&P Global

Over the last twenty five years, there have been three years where more than 50% of active funds have beaten the S&P 500 index, and barely so, and the extent of underperformance in the remaining years i staggering. The pushback from some active fund managers is that, given their investment styles, the S&P 500 is not the right index to use to judge them. Value fund managers, who invests in low-risk and high-dividend paying stock will argue that the portfolios they create are less risky than the S&P 500, making their lower returns more of a risk effect than underperformance. Academic studies that used risk and return models to tweak the index returns were quickly dismissed as being wrong, because the models that were used (the CAPM, the APM, Multi-factor models) were flawed. This deadlock was broken by S&P, when it created SPIVA, where the returns earned by fund managers in any class (small cap vs large cap, growth vs value, domestic vs foreign) are compared to returns that investors could have earned by investing in index funds in the same class, and by Morningstar, using a variant of the same approach. Thus, a fund manager who invested in small-cap, high dividend paying stocks would see his or her returns compares to the returns you would have earned on a small-cap, value index fund, making it much more difficult to explain away underperformance. There are many reasons for the slippage in active investing’s share of overall investing in the last two decades, but the SPIVA results are devastating and damning for any claim of active investing superiority. Here, for instance, are the results, by investment class, on the percentage of active fund managers who beat index funds in their investment style, over the last decade, at the end of 2025:

There is very little hope in these numbers, as fund managers underperform their respective indices in every single category, and by more, over longer periods. In fact, there is not a single fund group in any style that outperforms its respective index past ten years. For those of you who are reading this other geographies, and believing that it is different in your local markets, either because insiders have privileged access to information or market inefficiencies, SPIVA also tracks fund manager performance outside the United States, and reports similar results:

At this point, there is almost no counter to the argument that active investing collectively creates a drag on portfolio performance, and active investors seeking to defend the profession are left looking through the data entrails, hoping for niches where “alphas” exist. Thus, two decades ago, the notion that private equity investors and hedge funds were smarter than the market and could beat the market fueled a push of retirement and endowment money into these vehicles, but as they have become larger, they have come to resemble mutual funds, in terms of performance.

Investing Vehicles

There is a third use of indices that, in my view, has overwhelmed the measurement and performance evaluation roles that they play, and it is that they have become vehicles for investing in the form of index funds and exchange-traded funds (ETFs). That possibility was already existent in 1981, but at that time, the only index with an index fund available to most investors was the S&P 500. Today, you can not only invest in index funds across geographies, sectors or sub-groups based on fundamentals (including volatility, size and earnings), but the exchange traded fund explosion has given you an alternate route, with slightly higher costs (than index funds) and more liquidity.

Vanguard, a pioneer when it listed the S&P 500 index fund in 1976, now lists more than a hundred ETFs and more than two hundred index funds, allowing investors to not only invest in almost any market, but also in the subsets (sectors, small companies etc.) that they chose to. The size of the index fund business and the fees its creates for the index creators has created some tensions in the process, and faced with a choice between a poorly constructed index that index fund investors would love to trade on and a better constructed one that investors find less attractive, it is possible and perhaps even likely that the fund creators will pick the former.

Indexing and Passive Investing - Unintended Consequences

There are two forces that the last section highlight that have played out in altering financial markets structurally and dramatically in the last two decades. The first is that the underperformance of active investing has become easier to document and more visible for everyone to see. The second is that investors who see this underperformance have more passive investing vehicles in the form of index funds and ETFs accessible to them, and can move their money into them. As a consequence, in the battle between active and passive investing for investor dollars, the fight is getting so one-sided that, if you were a referee, you would invoke the mercy rule and try to stop it:

It is undeniable that the rise of passive investing vehicles has empowered investors, who have gained in terms of choice and costs, but it has also created an existential crisis for active investors, and especially so for those who make a living out of managing other people’s money. While some active investors hold on to hope - that they are better than the rest, that this is a cycle that will reverse, that some new technology (big data, AI) will save them- there are others who have taken a different tack. Mostly conceding that index funds and ETFs have outperformed fund managers, they have taken to arguing that the rise of passive investing is creating costs and effects that outweigh its benefits, and that action is needed urgently, though it is unclear from whom. In this section, I will highlight three of those effects.

1. The Index Inclusion Boost

The addition of a company to a well known or widely followed index (the S&P 500, Dow 30, MSCI Global) yields pluses, increasing its visibility to investors and potentially making it more investable, as well as increasing trading volume on the stock and making it more liquid. The question of how those benefits get priced in when a stock gets added to an index (and the costs of being removed from an index) have been studied over time, with a focus on additions to (and removals from) the S&P 500. As passive investing has grown, with more choices in index funds, it remains true that a significant percentage of passive investors hold S&P 500 index funds, and that, in the eyes of some, this should make inclusion in the S&P 500 index an even greater positive today than in decades past.

Looking across these studies, there seems to be a consensus that there is a bump up in the stock price from a company being added to the S&P 500, and bump down, when a company is deleted from the index, but disagreements both about the magnitude of the bump and whether it is permanent and transitory. The general sense that you get from studies is that bump in stock prices from being included in the index has become smaller and more transitory over time, and in the last decade or two decades, it has largely disappeared. To back this up, I look at one of the most complete studies that I have seen of the index inclusion question, where S&P took a look at the 715 companies added and 711 company deletions made to the S&P 500 between January 1995 and June 2021, and examine the excess returns in the days around the change:

Link to study

As you can see, the positiv eprice effects of being added to the S&P 500 index have depleted over time, as have the negative price effects of being removed from the index. Note that this finding cuts against the argument that as passive investing has increased in the last two decades, the allure of being in the S&P 500 should also have gone up. Instead, as the value of funds indexed to the S&P 500 has surged over the last two decades, the effect of being added to or taken out of the index has become smaller, not larger, and there is evidence accumulating that companies that get added to the S&P 500 are more likely to underperform than outperform in the twelve months after the addition.

As the debate about whether SpaceX, OpenAI and Anthropic should be included in the S&P 500 index heats up, I would suggest that the evidence of a small and dissipating price effect of inclusion has to become part of the discussion. I am sure that there will be some who will disagree with me, but I don’t think the price trajectories of any of these firms will be altered by whether they are included in or excluded from the S&P 500. For those who disagree with me, and believe that being added to index is bullish for investors in these companies, I would recommend that you look at the chart from this study which took a look at Tesla’s stock price behavior before, during and after its replacement of Apartment Investment and Management (AIV) in the S&P 500 on December 18, 2020.

Tesla not only under performed the S&P 500 in the months after its inclusion in the index, but massively underperformed the company (AIV) that it replaced in the index.

2. Momentum versus Fundamentals

As the funds invested in index funds and ETFs has surged, the argument that some are making is that being added to an index gives you a boost, largely because of index fund flows to companies in that index, and more of a boost if you are a large cap company. Since the money flows from other stocks, this line of thought also suggests that fundamentals will receive less attention and disconnect more from prices, especially because there are fewer active investors doing research and looking for market inefficiencies. In addition, they note that since indices are mostly market cap weighted, this momentum benefits larger market cap companies, in effect allowing them to become larger. As evidence in favor of this argument, they point to the fact that markets have become top heavy, where a few winners are carrying the entire market, that the small cap premium, an enduring feature of equity markets in the twentieth century, has largely disappeared in this one and the dominance of momentum in investing success in the last decade. I concede that these phenomena are consistent with the “passive investing feeds momentum” story, but I am skeptical that passive investing is the cause for the following reasons:

1. Momentum can cut in both directions: It is true that funds flowing into index funds flow into the companies in that index, with more flowing into large cap companies, but it is also true that funds can flow out of index funds, and when that happens, the momentum can cut in the other direction. In fact, the conclusion is that inclusion in a widely-tracked index (like the S&P 500) will increase intraday and short term volatility, but not price levels. In fact, the fading price bump from being added to the S&P 500 that we noted in the last section is an indication that the market does not buy into the momentum story.

2. Winner-take-all economics: If momentum is the reason for the big companies winning, there should be divergence between small and big companies on how fundamentals get priced. Put simply, you should see the pricing metrics (PE ratios, EV to EBITDA) for large cap companies increasing relative to small cap companies, as passive investing surges. Looking back at the fading small cap effect and top-heavy markets of the last decade or two, I would note that not all large cap companies have been winners, and the winning large cap companies have delivered a disproportionate portion of increased earnings. In the context of the Mag Seven, I have talked about how technology and disruption has changed more industries into winner-take-all businesses, with a few companies dominating these businesses, and why that phenomenon will play out in markets as well.

3. Active investing and equity research: In my view, the notion that most analysts and active investors are looking for market inefficiencies and seeking out information strikes me as misplaced. Much of active investing is built around publicly available information and a belief in the power of mean reversion, not original research and seeking information. It is true that there is a subset of active investors and equity research analysts who contribute to making prices more informative, but that subset is a small one, and one that is better equipped to survive the passive investing shift.

Doomsday stories about how passive investing is making markets less efficient and less inclined to reflect fundamentals strike me as overwrought, and while active investing will continue to lose market share, and deservedly so, it will not disappear. Since these stories are often being told by fund managers who not so long ago spoke contemptuously about efficient markets as an academic fever dream, they also strike me as both hypocritical and self-serving.

Even if you accept the argument that passive investing is making markets less efficient and more momentum-driven is true, I am unsure about the implications for investing. Asking individual investors, retirees and endowment funds to pay fees to professionals to manage their money while underperforming indices, in service to the larger cause of market efficiency is tone deaf and a non-starter. In fact, any endowment or pension fund manager who uses this argument to steer endowment funds to active money managers would be in violation of his or her fiduciary responsibility. Perhaps, the argument is being made to regulators to restrict index funds (on which indices they can index, how much money they can manage), I can see why active money managers may be in favor, because I understand that they are trying to protect their livelihood, but they should dispense with any talk about protecting individual investors or making markets more efficient.

3. Hidden risks

An undercurrent in some of the opinion pieces that I have read about why SpaceX should not be included in the S&P 500, written by investment experts and academics, is that it will expose retail investors and retirees to risks that they are unaware they are taking, or even if made aware of the fact, should not be taking in the first place. In particular, these opinion-writers seem to be arguing that the risks associated with investing in a big, money-losing companies (like SpaceX, and presumably OpenAI and Anthropic, when they go public) are so large that individual investors and retirees would not invest in these companies, and even they would, they should not be allowed to do so. I find this chain of reasoning to be both misguided and condescending, and reflective of misconceptions that are deeply and widely held in the investment expert class:

1. Risk and Diversification: Is it true that individual investors, if made aware of the companies that they owned in index funds, would blanch at the risks that they have exposed to in individual holdings? Investing just in a portfolio of a few companies like SpaceX would be imprudent, but an investor in a S&P 500 index fund is far less exposed to underperforming the market than the typical active money manager who either over invests in SpaceX (if it goes down) or chooses not to invest in it (if it goes up).

2. Smart and Stupid Money: While the investment experts and academics who push to protect retail investors and retirees from their own mistakes will never put into words this belief, implicit in this push is the view that these small investors are uninformed and naive, and will be exploited by smart money (institutional investors and hedge funds). The notion that the smart money will know whether SpaceX (and companies like it) is overvalued or under valued, and is positioned to time investments better is fanciful, since institutional investors are more traders than investors, making them market followers, not leader.

3. Good businesses and good investments: The weakest link in the argument against putting your money in money-losing companies is the implicit belief that money-making companies are good (safe) investments and that money-losing companies are bad (risky) ones. I will wager than an investor who was constrained to invest only in money-making businesses in the last two decades would have under performed an investor operating without those constraints, even after adjusting for risk. At the right price, a money-losing company can be a good investment and at the wrong price, a company with solid and stable profits can be a bad investment.

I am generally skeptical of attempts to protect individual or retail investors from their own mistakes and decisions, since more damage has been done to this group by those claiming try to help and protect them over time than by those who are out to exploit them.

Conclusion

In the week prior to the SpaceX IPO, S&P removed some of the suspense in the question of whether the company would be included in the index by announcing that they would stick with their requirement that a company be listed and traded at least a year before becoming eligible for index inclusion. That decision also means that OpenAI and Anthropic, if they do go public this year, will also have to wait a year for consideration. I am glad that S&P is not changing the rules to allow these companies to jump the queue to get into the index, but I hope that it is not framed as a decision that was taken to protect investors or in the hope that these companies would become magically money making, better governed and with working business models. The truth is that a year after they list and start trading, these three companies will still be money losing businesses, with business models that are still works in progress and will remain corporate governance horror stories. S&P needs the time to manage the transition of three trillion-dollar companies into the index, even as it confronts the challenge of claiming to be a large cap index that does not include three of the largest market cap stocks in the market. As for the companies (SpaceX, OpenAI and Anthropic), I will wager that they will lose little in market momentum from not being included in the index, and that their price paths will be determined by how the AI story continues to play out in terms of both substance (growth, unit economics, reinvestment) and perception (hype and momentum). The bottom line is that S&P needs these companies in its index more than they need to be in the index, with the consequence that the companies will not go out of their way to meet index requirements that they feel are costly to them, and that if there is any bending, it will be S&P that does it.

YouTube Video

The Prospectus: Data versus Information

The requirement that companies that plan to go public in the United States have to register with the Securities Exchange Commission (SEC) and file a prospectus has been in place for decades, but the contents have changed over time, with disclosures added on partly by regulation and partly in response to investor demands. In a paper focusing on IPO disclosures from a couple of years ago, I noted that prospectuses have become more bloated over time, often running four to five times longer than those filed by companies that went public three or four decades ago, but not necessarily more informative. The SpaceX prospectus that we made public on May 20, 2026, is 277 pages long, with an addendum that runs another 100 pages, with dozens of pictures (mostly of spaceships going into orbit), a soaring story, but with weak links and multiple distractions. To get a measure of how the prospectus changes my pre-prospectus story and valuation, I will start with the easy part of the update, where I use the numbers from the financial statements in the prospectus to replace my pre-prospectus estimates, on operating metrics like revenues and earnings as well as on share count and IPO proceeds. I will then move on to the weightier part of the analysis, where I assess how the information in the prospectus has changed my story line and value for the company.

The Prospectus: Data update

In my pre-prospectus valuation, where I assessed the value of SpaceX at roughly $1.2 trillion, I relied on scraps of information, including leaked stories of estimated revenues ($15.5 billion) and EBITDA of $8 billion, since I did not have access to the company’s full financial statements. With the release of the prospectus, that shortcoming has been remedied, and I started by updating the operating metrics that drive the intrinsic value of the company:

SpaceX prospectus

As you can see, my estimates for revenues for the launch and connectivity (Starlink) businesses were close to the actual numbers, but my xAI revenue estimates were much lower than reported. Overall, I had estimated an operating loss of $2 billion in 2025, and the prospectus yielded a larger loss of $2.57 billion. With almost $2 billion in interest expenses, unavailable prior to the prospectus, incorporated, the company reported a net loss of about $5 billion. A big factor in the operating losses reported by the company were its ballooning R&D expenses, and in keeping with my argument that these expenses should be capitalized, I estimated an earnings before interest, taxes and R&D of $4 billion.

On the financing front, the prospectus filled in details on cash and debt that were unavailable prior to the prospectus being made public:

SpaceX prospectus

My pre-prospectus estimate of book value of equity was a shot in the dark, at $20 billion, but the acquisition of xAI caused that number to jump to $41.3 billion, as did the total debt (inclusive of leases) to $22.9 billion. The former (book value of equity) played little role in my valuation, but ignoring debt of this magnitude may seem monumental, there are two offsetting factors that reduce the impact on my value estimate. The first is that I also ignored the presence of cash, and with $24.7 billion in cash, the company’s net debt is −$1.9 billion (cash exceeds debt), making the impact on value minimal. The second is that with my estimate enterprise value of $1.21 trillion, the debt, even if considered in full, is small enough to represent rounding error.

The prospectus did contain information on share count and structure, as well as on the company’s plans for the proceeds, and both were useful at the margin, with the former affecting my estimated value per share and the latter determining the treatment of the cash that will be raised from the offering:

• Share count: In my initial valuation, I used the private company pricing per share in conjunction with estimated market cap to back out a share count of 2467 million shares. With the prospectus, we get a clearer sense of shares outstanding, with a basic share count of 12,535 million shares reported in the prospectus (pages 246 & 247) in computing per share numbers. That share count does not include the new shares that will be issued in the offering, but that share count will be determined by the magnitude of the offering as well as the expected issuance price, and while the total share count includes options, warrants and rights that are exercisable before June 30, it does not include restricted stock units held by employees (see prospectus, page 18) and that information is still blanked out in the prospectus.

• Use of proceeds: It is estimated that SpaceX plans to raise $75 billion from the offering, and the prospectus specifies that the company plans to hold the proceeds to cover infrastructure investments in these businesses (see prospectus, page 66). That implies that any money raised in the offering will add to the company’s cash balance, right after the offering, and will augment firm value (but not enterprise value).

The prospectus also lays bare the governance questions that will overhang the firm, with information that there will be two classes of shares- 6,932 million class A shares with one vote per share and 5,602 million class B shares with ten votes per share. The public offering will be class A shares, and with Elon Musk holding all of the class B shares, he will control more than 85% of the voting rights in the company. In summary, the prospectus is long and filled with distractions, but there is almost nothing in it that surprises me. SpaceX is a growing company that is money-losing and cash-burning, that will be a Elon Musk vehicle (with all the pluses and minuses that entails).

The Prospectus: Story update

In my original post, I noted that SpaceX is a company, where it is the story about how its businesses will evolve over time that drives value, rather than the base year numbers (on revenues, earnings and cash flows). That story, broadly speaking, has three key spokes to it and they are summarized below:

The first of these spokes, target revenues, frame how big each business can grow over time, and is a function of the total market and market share. The second, the target operating margin, will capture how profitable each business can become, and is determined by unit economics and economies of scale. The third, reinvestment, measures how much each business has to invest to get to target revenues, and will vary with the capital intensity of the business. To frame how my valuation will change, as a result of what I learned from looking at the prospectus, I will start by presenting by pre-prospectus estimates on these key inputs, and then look at the impact of the prospectus on each input.

Pre-prospectus inputs and value

My pre-prospectus valuation of SpaceX contains my storyline for the three businesses that the company is in, with an add-on for the expansion options embedded in each business:

With these inputs in place, I estimated a value of $1.2 trillion the SpaceX enterprise, and since I ignored cash and debt, this yielded an equivalent market value. Driving these numbers are upbeat stories about each of the three businesses that SpaceX is in, with large revenues and high margins in stable growth.

The Prospectus Effect

To the extent that the prospectus contains information that alters the storylines on any or all of these businesses, it will affect my estimate of value for SpaceX.

1. Revenue Growth (Target Revenues)

I will start with the growth (target revenues) input and use two parts of the prospectus to reexamine my story. The first is the historical growth reported by the company for each of its three business lines - launch, connectivity and AI.

As you can see, the company saw its revenues grow by a third in 2025, relative to 2024, with divergence across businesses; the connectivity business led with revenues growing by almost 50%, the AI business saw an increase in revenues of about 22% but the space business reported only modest growth in the year (7.64%). In short, notwithstanding the star role played by AI and the appeal of the rockets in the space launch business, it is Starlink that carried the company in 2025. The prospectus mentions Colossus, xAI’s compute center, which has been leased to Anthropic for an eye-popping $1.25 billion a month, which should kickstart revenues next year, with the potential of tension in future years if xAI plans to go head-to-head against Anthropic in the AI products market.

The other relevant section of the prospectus contained estimates of total addressable market (TAM) for the company, broken down by business:

If the prospectus is to be believed, SpaceX has the largest TAM of any company in history, with a total TAM of $28 trillion, and AI accounts for $26 trillion of that market estimate. This estimate borders on fantasy, but I will cut the bankers who came up with these numbers some slack for two reasons. First, the estimation of TAM has been gamified by Silicon Valley, with bloated and patently unreachable numbers floated for companies, as I noted when I valued Uber (which was given a TAM of $5.7 trillion in its prospectus) for its IPO in 2019 and Airbnb (with a TAM of $3.4 trillion in its prospectus) in 2020. Second, it is true that AI agents are usable across almost every business and geography, giving it much wider reach than most products and services, and while the details of how the TAM was estimated are not specified in the prospectus, my guess is that the $26 trillion estimate includes all or most of the operating expenses of all businesses.

Story takeaway: I will stick with my estimates for target markets for the space launch and connectivity businesses, since the TAMs in the prospectus are, in my view, over reaches, and I will slow growth in the near years, to reflect that these businesses will take time to mature. In the AI business, I disagree with the magnitude of the TAM in the prospectus, but the acquisition of Cursor and the indications in the prospectus suggest that xAI very much wants to be part of the enterprise solutions space, notwithstanding its immense capitalization needs, and I will double my target revenues for AI from $80 billion to $160 billion, reflecting my estimate of a TAM of about $3 trillion to $4 trillion for AI products and services from businesses.

2. Profitability

On the profitability front, the first part of the prospectus that I looked at was its breakdown of income statements, by business:

With the caveat that we have only two years of detailed information, there are interesting findings that emerge from the historical data on each of the businesses.

• The space business has the best unit economics of the three business, with a gross margin of about 67%, reflecting the cost advantages of its reusable rocket technology. While the space business reported an operating loss, that was entirely because of its weighty R&D expenses, and capitalizing those expenses results in a healthy operating margin for the business.

• The connectivity business does not have gross margins as high as the space business, but those gross margins are improving, with gross margins jumping from 37% in 2024 to 48% in 2025. This business had positive operating income in 2025, even before capitalizing R&D, and improves substantially with capitalization.

• The AI business not only has the lowest gross margins of the three businesses, but saw deterioration of those margins in 2025, reflecting intense competition from other LLMs as well as the rising costs of delivering AI products and services.

There are other parts of the prospectus that come into play in the profitability discussion, with each of the businesses:

• On the space launch business, the cost of launching payloads at SpaceX have been trending down, making its already large cost advantages in the business even larger.

• On the connectivity businesses, there is bad news and good news on the per user front. The bad news is that the revenues, per month, per subscriber, declined from $99 in monthly revenues in 2024 to $66 in monthly revenues in the first quarter of 2026. The good news is that the number of subscribers has doubled from 5 million in the first quarter of 2025 to 10.3 million in the first quarter of 2026, with the bonus that the company has been able to improve its profitability (see gross margins in the table above) over time.

• On the AI business, there is not much to go on, on the profitability front, since the focus in the prospectus is more on the increase in compute capacity (see nameplate compute draw on Page 90 of the prospectus) than it is on revenues, especially on the enterprise front. Here again, though, the Colossus lease with Anthropic should help with profitability in the near term.

Story takeaway: The unit economics for the space businesses, in conjunction with the recognition that there are no other substantial operating expenses (outside of the misclassified R&D expense) in either business, lead me to increase my estimate of the target margin for the business to 45%, from 40%. I will leave intact the target margin of 60% for the connectivity business, because once the satellites that service this business are in space, this is the business that will benefit the most from scale. My biggest shift is in my estimated target margin is for the AI business, where the dynamics that are pushing gross margins down, i.e., increased competition and high costs of delivering AI services, will persist; my estimated operating margin drops from 45% to 25%.

3. Reinvestment

In my post prior to accessing the prospectus, I did describe SpaceX as a capital intensive business, but the actual spending on capital expenditures and R&D in the prospectus is breathtaking in its magnitude:

In 2025, the company spent almost $14 billion in capital expenditures and almost $9 billion in R&D, a doubling of its reinvestment from 2024. In particular, it is AI that is driving the bulk of this surge, accounting for more than $14 billion in total reinvestment in 2025, with $9.1 billion in capital expenditures and $5.1 billion in R&D. The positive twist that a SpaceX optimist would put on these numbers is that the spending on AI in particular is a positive, indicating that the company is not planning to settle on a niche market strategy, but instead will will go head-to-head with Anthropic, Google and OpenAI for the enterprise solutions markets. The negative spin is that this ambitious agenda will translate into tens of billions more in capital expenditures in the near years, creating a drag on the cash flows and value destruction if they lose the AI market competition.

Story takeaway: Given that SpaceX is continuing to invest substantial amounts in its space launch and connectivity businesses, I will increase reinvestment in the near term (years 1-5) by lowering how much they will generate as additional revenues for every additional dollar of capital invested (lower sales to capital ratios). With AI, where I was already assuming that reinvestment would be large (with a low sales to capital ratio), the tripling of target revenues will result in a surge in reinvestment to generate the higher sales.

Updating Story and Value

While the core story of SpaceX being a company with growth potential and strong competitive advantages that I framed prior to reading the prospectus remains intact, there are changes to that story that come from the information in the prospectus. The prospectus reinforces the notions that the company is best positioned in the connectivity business to generate both revenue growth and profits in the near term, that its cost advantages in the space launch business will persist and deliver profits, but that target market will be slower to develop, and that the AI business has both the largest target market and poses the biggest challenges, in terms of profitability and capital intensity, for SpaceX.

Bringing together my changes in target revenues, operating margins and reinvestment inputs allows for an update of the input table that I started this section with:

Clearly, some of the changes in inputs (such as the higher margins for the space launch business and a bigger target market for AI) will push value higher, and some of the inputs (including a slowing of near term growth for all business, and the much lower margin for the AI business) will push in the opposite direction. Since US treasury rates have risen from 4.20% at the time of my earlier valuation to 4.56% at the start of June, I have increased the costs of capital that I use in the valuation accordingly (to 8.37% from 8.02% to start the valuation, and the steady state cost of capital to 8.25% from 8.00%; both numbers would put SpaceX at close to the median for all US companies). With these updated inputs, I reestimate the cash flows and the valuation for SpaceX, with the IPO proceeds (estimated at $75 billion) added to the mix:

Download spreadsheet

The enterprise value for SpaceX edges up from $1.21 trillion, in my pre-prospectus valuation, to $1.22 trillion with the post-prospectus numbers, and the overall equity value increases to $1.3 trillion, with almost all of the increase coming from the influx of $75 billion in cash from the IPO, albeit with a higher share count. The value per share of about $100 will need some revisiting as the IPO numbers firm up and more information is forthcoming on restricted stock units owned by employees, but just as I was finishing this post, a news story hit the wires that the offering price would be set at $135/share.

If I were to summarize the impact of the prospectus on my SpaceX story, it would be that it has made the story bigger, but also more volatile. There are a multitude of risks that SpaceX faces in each of its businesses, but the one that I would be concerned about the most is that it will overreach in the AI business, beginning with an overestimate of the target market for AI products and services and the strength of its own competitive position in that market, and following through with investments that reflect those misplaced assessments. Those concerns are heightened by a voting share structure that locks in Elon Musk’s control of the company, since there is little that shareholders can do to restrain the company, if SpaceX doubles down on capital expenditures and acquisitions in the AI space, even after it becomes clear that the AI market is much smaller than anticipated and/or that xAI’s offerings are not as good as the competition. If you add to this mix the antipathy that exists between Musk and Sam Altman, you have the potential for a UFC match between two monstrous egos, funded by tens of billions of dollars shareholder money.

Financial Statements and Value: The Life Cycle Effect

Financial analysis and valuation, going back to Ben Graham’s Security Analysis, has always been centered on financial statements, and that focus has become more intense over the last few decades as access to data and analysis tools has expanded. In fact, much of what passes for valuation has become financial modeling, where line items in financial statements are forecast based upon the historical time series, with the proverbial bottom lines being earnings and cash flows. Along the way, ratios computed from financial statement numbers are used to screen companies for investment quality. Some of these ratios, such as accounting returns on capital and equity, have become the basis for assessing company quality and competitive moats in the hands of consultants and investors. The SpaceX prospectus is a case study in why this approach to investing is often myopic and misleading, and why the informational value of financial statements will change as companies grow and mature.

In valuing companies, you are always trying to forecast revenues, profits and cash flows in future, but they key questions you want answered and the drivers of value shift as you move through the life cycle:

As you can see, for young companies, the key determinants of value include sizing the total market and assessing unit economics, and not the proverbial bottom lines in accounting statements including the magnitude of revenues and profitability. As companies move through the life cycle from start-up to mature to decline, you should expect to see financial statements evolve as well. Young and high growth companies will generally report small revenues (though they expect those revenues to ramp up over time) and losing money and having negative cash flows is a feature, not a bug. As companies mature, revenues will get larger (albeit with lower growth) and profits turn positive, as will free cash flows available to return to shareholders in dividends and buybacks.

If you allow for the fact that all three of SpaceX’s businesses are young, falling in the young to high growth categories, the big questions driving value are about market size and unit economics, since the former provides the basis for revenue growth and the latter determines profitability. That is why, when looking at the prospectus for SpaceX it was the data on total addressable markets, unit economics and capital intensity that had a bigger impact on value, and this information, for the most part, was in the footnotes to the financials, rather than in the financial statements themselves.

For those who are focused on value metrics/constraints (consistently money making, high profit margins and accounting returns) and pricing multiples (low EV to EBITDA or low PE), the SpaceX prospectus is full of red flags. SpaceX is a company with small revenues and large losses, and paying a hundred times revenues for it (which is where a $1.8 trillion pricing would put it) seems foolhardy. I have no quarrels with this point of view, which animates old-time value investing, but this perspective comes with a cost in terms of investment choices. Investors who are wedded to never buying money losing companies or never paying more than twenty times earnings for a stock will end up with portfolios of mature (and declining) businesses. If that is their comfort zone, the strategy is perfectly defensible, but they should dispense with complaints about never being able to find high growth stocks to invest in or critiques of others who find these stocks attractive, notwithstanding the weak numbers.

There are many good arguments that can be made about why you should not invest in SpaceX, but basing that conclusion on the fact that they are money-losing or have negative cash flows or trade at a high multiple of revenues is both lazy and unconvincing. In contrast, making a case against investing in SpaceX because you believe that the target markets for its businesses will be far smaller than the company thinks they will be, or that cost and competitive pressures will drive margins down or even that you find its corporate governance structure and dependence on a personality (Elon Musk) off-putting is perfectly reasonable. If you do make that case, though, it is worth remembering that this is your point of view, and that disagreements about market size and profitability across investors, especially in young companies, are natural and healthy. In short, based on my inputs and story, I think that SpaceX is worth about $1.25-$1.3 trillion, but if you contend that it is worth $3 trillion or only half a trillion, it is neither my job nor my place to convince you that I am right and that you are wrong.

The IPO Pricing Game

In the coming weeks, you will undoubtedly be exposed to multiple perspectives on SpaceX, and that is healthy. That said, you will be better equipped to make sense of these perspectives, and perhaps incorporate some of the views into your own, and reject those that do not make sense, if you have an understanding of what an IPO process involves. In particular, understanding the motivations of the different players in the game (the investment bankers setting the offering price and managing the offering, the issuing company, the investors and traders jockeying for shares at that offering price and the traders positioning themselves for the first day of trading) will help determine whether you should be playing this game or sitting it out, at least for the moment.

The Bankers

Let’s start with the sequencing that goes into a conventional initial public offering, though alternatives have emerged to it in recent years:

As you look at the role played by bankers to the IPO process, allowing them to keep a slice of the IPO proceeds, the SpaceX IPO is a testimonial to the dwindling value added by bankers on every dimension:

1. Timing: It is urban (or market) legend that investment banks can time markets, and that this market timing can help determine the best time to go public. Just one look at the track record of market strategists at investment banks should dispense with this delusion, since banks (and most institutional investors) are (and have never been) good at gauging market momentum and shifts in mood.

2. Filing and Offering details: It is true that there are technical details and logistical steps to filing a prospectus and setting offering details, but they are almost all mechanical. With SpaceX, I am not sure whether the prospectus, as filed, was the work of a team of bankers, but if it was, I wonder what an entirely Grok-written prospectus would have looked like, and whether we would have noticed the difference.

3. Pricing: In an IPO, the bankers’ mission is to price companies for their offering, not value them, and while they usually draw on pricing multiples and peer groups to make that pricing judgment, they are guided by the pricing in the most recent private transactions, usually in the form on venture capital rounds. With SpaceX, that task is simplified by the reality that this company, while private, has had active trading in its private shares, and that it was priced at roughly $1.2 trillion prior to the IPO process commencing. Adding the $75 billion in offering proceeds, and incorporating the advantages of increased liquidity from being a public company and becoming part of the S&P 500, it is not surprising that there is a sense that the offering will be priced at between $1.5 trillion to $2 trillion, with or without the investment banking input. My guess is that we will end up somewhere in the middle, with some handwaving about revenue multiples and other AI companies used to justify that pricing. (After I finished this post, a news story popped up that the offer price would be set at $135/share, translating into about a $1.8 trillion pricing for the company.)

4. Selling/Marketing: In an age where investment banks have lost credibility and social media is where marketing happens, SpaceX can generate its own marketing spin, and has an army of influencers behind it. In addition, almost every institutional investor has a point of view on whether to own SpaceX or not, it is unclear what exactly a roadyshow can do to augment the sales pitch.

5. Price guarantee: The pricing guarantee that investment bankers offer in initial public offerings is a mostly empty promise, since they systematically set offering prices at below (by 15-20%) what they believe the market will pay. That is the reason that the offer price for SpaceX will be set below the upper end of the range, and while the discount may seem like a significant loss to funders and current owners, the fact that the offering is for less than a tenth of the shares in the company will soften the blow.

6. Post-market support: As a follow-up to the price guarantee, investment banks often offer after-market support for companies in the days after they go public, buying shares if the stock comes under selling pressure. With SpaceX, that option is off the table, since no investment bank has the capital to support the pricing of a two-trillion company, if investors turn negative on it.

In fact, given that banks are perhaps getting more from the initial public offering, in terms of publicity and allotments for their preferred clientele, than SpaceX is getting from their services, you could argue that the bankers should be paying the company for reflected glory, rather than charging them fees. The only good reason that I can think of for SpaceX not going the direct listing route, where you dispense with the kabuki dance of offerings and let the market set the offering price, is that the company needs the cash from the offering, and that route is much more difficult to take in a direct listing.

Issuer (Company, Founder and Investors)

Looking at the IPO from the SpaceX perspective, the public offering will provide benefits. For the investors in the company in its private form, including venture capitalists from early in its life to public investors in more recent years, the IPO will allow them to cash out, albeit after the lock-out period expires in a few months. For the company, the increased access to capital from being a public company will allow it to fund the capital expenditures and investment needs that emanate from the company’s ambitions in the AI business. For Elon Musk, the public offering has the potential to make him the first trillionaire in history, in addition to unlocking new pathways to further enrichment for meeting specified targets (including getting a million people on Mars).

Since some of these benefits have been in existence for many years, the fact that company stayed private for that period is an indication that there are costs to going public that have held it back. The first is that, notwithstanding Musk’s voting control of the company, become a public company will open SpaceX to market scrutiny, in the form of earnings reports every quarter and insider trading reports. The second is that the market is fickle, and while it is rewarding companies that invest in AI with high market prices today, it can change its mind and punish them for the same reason. The third is that while there is little that investors can do to trade and make money on overpriced private businesses, they can sell short on public companies.

Investors and Traders

SpaceX is a company that has been in the public eye for a decade or more, even as a privately owned enterprise, partly because of its social media boosters and partly because its space launches make it a magnet for attention. There are many who are drawn to the company, but unable to invest in it as a private business, will now have a chance to do so, if it goes public. But should they try to partake in the initial offering? The answer to that question depends on whether you are an investor, where you buy (sell) companies that you believe are trading under (over) their assessment of value and hope the gap closes or a trader, where you buy (sell) companies where you expect prices to go up (down) in the future, for a multitude of reasons, only some of which may relate to company fundamentals.

I am more investor than trader, and I say that without judgment, since the end game in markets is to make money, not score intellectual points. The truth is that I am not a very good trader, and I am better off staying in my preferred domain, which is valuation, albeit with no guarantees of a payoff. My valuation of SpaceX was driven by my interest in the company and belief that it is in unique, cutting-edge businesses, and my decision on whether to buy into the offering is therefore driven by my assessment of its value. At the rumored pricing of $1.8 trillion for the company, it is too richly priced for my tastes, given my valuation of $1.25-$1.35 trillion for the equity in the company. That does not mean that I will never buy the stock, since the market does change its mind, and if the price does drop by enough, my decision would change accordingly. It is worth remembering that Facebook was selling at half its offering price a few months after its IPO, and that Uber lost more than 50% of its market cap in the year after its public offering, moving both companies from over to under valued.

If you are a trader, though, the game changes. Specifically, the intrinsic value of the company is not central to your decision, perhaps even irrelevant, and your judgment on whether you seek to partake in the SpaceX offering will depend on your reading of market mood and momentum. I would not be surprised in the least to see the offering priced at $1.8 trillion, and see a jump in the price on the day of or in the weeks after the offering, and if that is your most likely scenario, being able to get into the offering at the offer price or even in the first few hours or days of trading will be a winning strategy. The risk, of course, is that momentum can shift quickly, causing a significant price drop, effectively making timing your trades right key to your trading strategy. The shifting and often unpredictable forces of mood and momentum are also the reason that as an investor, I would not sell short, notwithstanding my value assessment, even if the pricing for the company pushes from $1.8 trillion to $2 trillion or more.

A Loaded Bet on AI!

As the IPO process for SpaceX heats up in the coming weeks, you should prepare yourself for a flood of selling from the company and its bankers, with talk of possibilities and potential dominating the discussion, as well as arguments from the other side, where it will be framed as a vehicle for AI hype, destined to fail. If you are on the receiving end of these sales pitches, you should listen but check the numbers for plausibility and make your own judgments. For the bankers involved and the issuing company, the biggest danger to a successful offering is not that there will be near-term reality checks on their hype, but that the market mood will shift, either in the aggregate or specifically related to AI, in the weeks leading up to the offering. No matter what your views are about the SpaceX IPO, positive or negative, there is no denying that this company is a loaded bet on the AI and Elon Musk, and while that may concern some, there are others who will look at Musk’s track record with Tesla and feel the odds are in their favor.

YouTube

Attachments

1. SpaceX prospectus

2. Valuation of SpaceX, post-prospectus on 6/2/26

Blog posts on SpaceX

1. To a Trillion(s) and Beyond: A SpaceX Odyssey!

Apple’s CEOs: From Scott to Jobs to Cook!

Unfair though this may seem, the Tim Cook story at Apple has to start with Steve Jobs. Jobs co-founded the company in 1976, with Steve Wozniak, and while the company went through a series of CEOs in the next two decades, Jobs was the face of the company in its early years. While it is easy, with the benefit of hindsight, to view these as good years for the company, those early years reflected both Job’s strengths and weaknesses. His vision and force of personality gave rise to the personal computer in its current form, as a tool for everyone to use, not just tech geeks, and as someone who bought his first Mac (the 128K without a hard drive) in 1984, and has stayed a Mac user since, I am grateful. That said, the dark side of Jobs, manifested in impatience with underlings and an obstinate belief that he knew what customers needed better than they did, led to the Lisa, the only Mac I regretted buying almost immediately after my purchase, and a loss of business markets to Microsoft. Those failures led Apple to the brink of failure, and to Jobs being cast out of the company by its board in 1985, though the CEOs that followed had neither the strategic vision nor the business-building capacity to rescue the company.

In 1997, Apple looked like it was a company heading into oblivion, as Windows became the dominant operating system for personal computers, and it seemed like Apple had lost its purpose. The August 1997 return of Steve Jobs,, who had used his years in the wilderness to build Pixar, a company that revolutionized animated movie making, is now the stuff of legend, as he rebuilt Apple in the ensuing years into a powerhouse, around the iPod, the iPad and most of all the iPhone. While there are books and movies chronicling the Steve Jobs success story, it is worth asking what the difference was between the first iteration of Steve Jobs at Apple (from founding to leaving in 1985), where Apple lost ground to Microsoft, and the second iteration of Steve Jobs (from his return in late 1997 until his resignation in 2011). The first was that he was older, and to the extent that with age comes some wisdom, it helped, but it is unlikely to have been the change maker. The second was that in his period away from Apple, Jobs created and built up other companies, with Pixar being the biggest, where he learned to deal with people better and perhaps compromise a bit more than he used to. The third was that he benefited from the presence of Tim Cook, first as an executive in Apple sales and operations, and more importantly, as chief operating officer (COO) for Apple, starting in 2005. If Steve’s skill was vision, where he showcased Apple’s next “big innovation” at meetings in his trademark black turtleneck, Cook’s skill was building manufacturing hubs and supply chains to convert the vision to products. That separation of vision from business building created the Apple juggernaut in the first decade of this century. While that division of labor clearly was in the company’s best interests, Jobs deserves credit for being willing to set his ego aside and delegate the powers to make it happen.

Tim Cook has been CEO for fifteen years, and when he retires on September 1, 2026, he will have been the longest serving CEO at Apple. It cannot have been easy, especially in the early years, as the comparisons to Steve Jobs were front and center, and there was pressure on him to continue in the same path. To Cook’s credit, he never tried to be Jobs, and he created a very different template for himself, one that fit him and the company well, and served as a testimonial to his self assurance. In one of my talks about a decade ago about Apple, I described Cook, perhaps harshly, as a man without a visionary bone in his body, but one who would make sure that the trains ran on time (or the iPhones were delivered as promised), and I think that he has used that strength to good effect during his years as CEO of the company.

Apple’s Finances in the Twenty First Century: The Steve Jobs and Tim Cook Years!

Steve Jobs handed over a company to Tim Cook in 2011, that was extraordinarily profitable, and at the time of his leaving, already the largest market cap company in the world. While that fact leads some to discount what Cook has done at Apple since, I think it is worth going back in history and looking at corporate handoffs of great companies, and how often they become tangled messes, as new CEOs overreach and overpromise.

The place to begin our comparison of the Jobs and Cook tenures is by charting Apple’s market capitalization, with the delineation into the Jobs years (1998-2011) and the Cook years (2012-2026):

Looking across the aggregated years across both CEOs, it has been an extraordinary time. Apple began the Jobs tenure as CEO with a market cap of $1.68 billion, and by the end of 2025, its market cap had risen to over $4 trillion, and its performance burnishes the reputations of both Jobs and Cook. Jobs provided the foundational boost for the company and the innovations he presided over delivered a compounded annual price appreciation of 47.19% between 1997 and 2011, a period when US equities were struggling and Apple reached the top of the market cap table in 2011. With Tim Cook at the helm, Apple added an astounding $3.64 trillion in market cap, but a strong equity market provided strong tailwinds, and the company’s annual returns were more modest. On a percentage return basis, the Jobs years were better, but in my view, the fact that the annual returns in the Cook years were just as impressive, because they had to be earned on a much larger firm.

The reasons for Apple’s sustained increase in market capitalization were simple - solid revenue growth and a profit machine that delivered high margins, even as the company scaled up:

As with the market capitalization comparisons, this chart yields metrics that are favorable to both Jobs and Cook. Under Jobs, the company scaled up its revenues significantly, with a compounded annual revenue growth rate of 23.67% between 1997 and 2011, and just as significantly, went from posting subpar margins and a net loss in 1997 to becoming one of the most profitable tech companies in the world, Under Cook, revenue growth rates came down (to a compounded annual average of 8.52% between 2012 and 2025), but on a much larger scale, and the company preserved and grew its profit margins.

There was one corporate finance dimension on which Cook deviated from Jobs, and that was on cash return or dividend policy. In the chart below, I look at the cash returned to shareholders by Apple during the tenures of the two CEOs:

During Job’s tenure at Apple, the company paid no dividends and initiated only modest cash buybacks, mostly to cover stock-based compensations. With Tim Cook as CEO, Apple was one of the greatest corporate cash success stories of all time, initiating dividends in 2012 and increasing them over time, and supplementing those dividends with cash buybacks that, in the aggregate, were the largest in corporate history. In sum, the company has bought back almost $800 billion between 2012 and 2025, and the most astonishing feature was that, while returning all of this cash, the company also accumulated one of the largest corporate cash balances in history.

To the question of how Apple was able to return this much cash, increase its cash balance and still grow itself, the answers are three fold. The first is that the iPhone, perhaps the most valuable single product in business history, continued to deliver for the company, with modest reinvestment needed on its upgrades.

While much of the credit for the iPhone is still given to Steve Jobs, and rightly so for fostering the innovation, credit is also due to Cook, who has taken the franchise handed to him, and grown it on steroids. The second is that the company borrowed $17 billion in 2013, a Cook departure from a Jobs practice of avoiding debt, and it has added to that debt load over time, though it remains a small slice of overall value:

While much is made of Apple’s debt foray, it is worth recognizing that Apple is still a very lightly indebted company on any debt metric, and that if you net the company’s considerable cash balance out against its total debt, its net debt has always been negative (cash exceeds debt). In fact, Apple’s use of debt is so light that the only rationale for its existence is creating a presence in the bond market, just in case it needs to use it more in the future. The third feature is that the company has been cautious in its forays into new products and markets, especially outside its domain, and this shows up in two data series.

• The first is that while Apple has acquired more than a hundred companies, almost all of them are small, private technology companies with small price tags, with the intent being bringing their products and services into the Apple ecosystem after the acquisition. In fact, its largest acquisitions during this century are so small that they represent petty cash, relative to its cash balance as a company. Beats, for instance, which was one of Apple’s biggest acquisitions cost the company about $3 billion, a number dwarfed by its cash balance that year, which was more than $100 billion.

• The second is that in the last five years, as big tech companies have gone on an AI capital expenditure binge, Apple has been the outlier, holding back on its AI investments, and this can be seen in the chart below, where I compare Apple’s capital expenditures to those of the rest of the Mag Seven:

  

  As the rest of the group has ramped up its capital investments, with much of it going into AI, Apple has held back, and its share of the total cap ex at the companies has fallen from 8.04% to 3.02% over the period.

In sum, looking at the changes at Apple over the last fifteen years, the company has changed from the growth engine, driven by disruptions, in the Jobs years to a mature, cash-returning and more cautious company under Cook. I have posted more about Apple than about any other company in the world (and I have a sampling of some of those posts at the end of this post) and have been a shareholder in the company for significant portions of both the Jobs and Cook tenures. I have not always agreed with either man, on choices that they have made at the company, but I respected both of them enough to view them as good stewards of my investment. In a world full of CEOs who are quick to herd to what the consensus view is, I admire both men for their willingness to stand on their beliefs.

Vision or Restraint: A Life Cycle Perspective

If you were to create a profile of Tim Cook, the manager, based upon the choices that he has made at Apple during his tenure as CEO, two very divergent views emerge. To his admirers, his actions on some fronts (initiating dividends, massive stock buybacks, borrowing money) and inaction on other fronts (no big acquisitions, diffidence on AI investments), represent an exercise in discipline and restraint, preserving the company’s crown jewel (the iPhone) and fending off the bankers and consultants, with their false promises. To his critics, and there are quite a few, Cook’s caution has cost Apple its disruptor status, when it could have used its ample cash reserves to buy its way or invest in into almost every new business that has bloomed in the last fifteen years. In fact, they point to chances that Apple has had to buy some of the biggest stars in the market, from Tesla and Netflix more than a decade ago to Anthropic, Mistral and Perplexity in more recent years.

It is impossible to argue that one side is right and the other side wrong, but it is undeniable that both pathways (the restrained pathway that Apple adopted and the more aggressive pathway that it could have taken) include trade offs. It is true that Apple’s restraint has led it to miss out on some of the biggest trends in technology over the last decade, but it has also avoided the overpayment that is so common with high profile acquisitions of big companies. The argument that Apple would be worth a lot more today if it had bought Netflix or Tesla a decade ago falls flat for two reasons. The first is the selection bias in picking two companies that, in hindsight, have emerged as winners, when in fact there were at least a dozen other worse-performing companies that were also on Apple’s radar. The second is the presumption that companies like Tesla or Netflix would have been just as successful, owned by Apple, as they were as stand alone enterprises. The clash of corporate cultures that would have ensued if Apple had bought either Tesla, a company that reinvents its business narrative every few hours, or Netflix, an entity that makes content in quantity with the hope that some it sticks, would have been epic, with the risk that both Apple and its acquired target would have gone down in flames.

More generally, though, the question of whether you want a visionary or a disciplined business builder at the top of a firm is not one that has an easy answer, since it depends on the firm in question. In my work on corporate life cycles, I focus on the management skills that are needed most in a company, based upon where it is the life cycle, and that may help address the choice between vision and restraint:

With young companies, vision dominates, as managers work to sway investors, employees and nascent customers that their product or service will find a market. As the vision takes hold, converting it into commercial products and services requires trading off some portions of vision for pragmatism, in the interest of getting the business going. As products and services find demand among customers, business building becomes a key difference-maker, with the grunt work of marketing, production facilities and supply chains coming into play. Assuming that you have made it through these three stages, the trade offs of scaling up come into focus, and as you hit market limits, success depends on being opportunistic in finding new products and markets, but only if they exist. In corporate middle age, pathways to easy growth, especially at scale, become difficult to find, and to the extent that value comes from moats and core products, playing defense against competitors takes priority. Finally, in decline, a phase that no company ever wants to enter, but is inevitable at some point, you need to be willing to shrink a firm, shutting down businesses that no longer deliver value and selling other assets to high bidders.

Given these very divergent management functions, it should come as no surprise that there is no prototype for the perfect CEO, McKinsey and Harvard Business School blueprints notwithstanding. Viewed in this framework, I would argue that Apple has been lucky with its last two CEOs, both in terms of persona and in terms of sequence. When Steve Jobs rejoined Apple in 1997, the company had hit rock bottom, and with little to offer in liquidation, his vision allowed for a reincarnation, with disruptions leading the way, and as we noted earlier in this post, having a strong chief operating officer in Tim Cook made the difference. The Apple that Tim Cook inherited, when he became CEO, was a very different entity, already the world’s largest market cap company, with a superlative franchise in the iPhone. In corporate life cycle terms, Apple was a mature growth company, and what Cook lacked in opportunism , he made up for by defending Apple’s biggest product line(iPhone) and augmenting value with increments like the app store and devices. That said, while each of these men created value for shareholders, I don’t think that either would be regarded as highly, if you swapped their tenures in terms of timing. I don’t think Tim Cook would have been able to bring Apple from its near-demise to being on top of the corporate universe, if he had become CEO in 1997, and I think Steve Jobs would have been ill-suited to the Apple that was in existence in 2011.

Aging, Management Mismatches and Corporate Governance

In a post from a few years ago, I used the connection between CEO type and corporate lifecycle to examine why management mismatches occur at firms, and the consequences of that mismatch. Specifically, there are three confounding factors that can make matching up CEO to company, given where it is in the life cycle, complicated:

1. Like humans, companies age, but unlike humans, the rates at which different companies go through the life cycle can be wildly different. An infrastructure or manufacturing company can take decades to become operational, followed by extended phases of growth and maturity, before going into decline. In contrast, a tech company can have explosive growth early in its life, spend a brief period enjoying the fruits of its success as a mature company before declining precipitously. As a consequence, managers and investors who use chronological age as their corporate aging metric can misjudge where they are on the life cycle.

2. While aging is inevitable for both humans and businesses, some mature or even declining businesses can find pathways, either through happenstance or management choices, to rediscover their youth. These businesses become the stuff of legend, and they are the subjects of books and business school case studies, and their CEOs are elevated to management deities.

3. The narratives built around companies that reincarnate and the CEOs atop these companies also feed into management incentives and behavior. The story of Steve Jobs at Apple has been told and retold, but it is worth remembering that for every story of reincarnation, there are a hundred stories you can tell about other CEOs who tried to follow the Apple playbook, spending billions on reinventing their companies, with little to show in terms of payoffs. (See my posts on Marissa Mayer at Yahoo! and on Blackberry.) In essence, the glorification of CEOs who bet big on turnarounds at mature or declining companies, and win, sets up CEOs facing similar circumstances to behave like riverboat gamblers, when making management choices at their firms. After all, if their bets pay off, they join the legend crowd, and if they do not, they contend that they did their best, and that circumstances conspired to bring them down.

The bottom line is that there are a number of ways in which you can end up with CEO mismatches - a CEO who cannot adapt to the changing demands of an aging business, a hiring mistake or even changes in the macro environment, and when those mismatches occur, is is inevitable that there will be friction between the CEO and shareholders. In a sense, almost all corporate governance challenges can be traced back to management mismatches, and the power (or the absence of it) that shareholders have to fix those mismatches:

While Tim Cook’s time as CEO of Apple is now seen through rose-colored lens, it is worth remembering that Apple was targeted repeated early in his tenure by activist investors. While some of the changes that these activists were pushing for were warranted, some were not, and Cook deserves credit for not capitulating. Carl Icahn, for instance, wanted Apple to increase its debt substantially, borrowing hundreds of billions, but I took issue with his argument that Apple could borrow this money at the low rates that he was extrapolating. A couple of years later, David Einhorn made his play, arguing that Apple should issue preferred shares with a 4% dividend yield, and I noted that preferred stock could bring with it all of the cashflow commitments of debt, with none of the tax advantages. I have long argued that the best defense a management has against activist investors is delivering superior performance and returns, and Tim Cook delivered on both dimensions, and faced little more than sniping from disgruntled investors in his later years as CEO. In the last four years, the criticism has come primarily from analysts who fault his caution, and argue that Apple risks falling behind its more aggressive competitors in the AI race, but here again, Cook has stood his ground.

Management Transitions, Past and Present - The Mag Seven

I don’t envy John Ternus, who is Cook’s heir apparent, because he is following two CEOs who were immensely successful, albeit in different ways. If there are lessons he can learn from both Jobs and Cook, they include the following:

• Find your own path: There will be pressure from some investors to be just like Jobs, and go for big disruptions, or from others to imitate Tim Cook, and leave Apple as a cash machine. While it may take time, Ternus has to find his own path as CEO, based on not only what he brings to the table, given his background in computer hardware, but on what Apple’s strengths are as a company in 2026 and the markets it is facing right now.

• Adapt to the company you are managing: Just as the Apple that Jobs took over in 1997 was very different from the Apple that he handed over to Cook in 2011, the company that Ternus takes over is different from the ones handed over in either of the prior iterations. When you are at the helm of one of the largest market cap companies in the world, you have to start with the recognition that any new product or service that you introduce will have to be huge to make a dent in the operating metrics (revenues and profits) or market capitalization. In addition, the franchise that holds up the company’s cash machine is the iPhone, and Ternus cannot afford to take his eyes of that prize.

• Keep the feedback loop open: When you are a company worth trillions, with legions of shareholders, and hundreds of analysts, you will have advice meted to you constantly on what you should or should not do. Much of that advice will be bad, and should be dismissed, but some of it is worth listening to and perhaps converted into policy. In addition, as CEO, I hope that Ternus views the market price as a crowd judgment on Apple’s actions rather than the product of speculation, and accepts that while that judgment can be wrong, it should be taken seriously.

I wish Mr. Ternus the best, for purely selfish reasons. As a Mac and Apple device user, I want the company to prosper and continue to make products that I can continue to use on a daily basis, and as a shareholder, I want my investment to do well.

The attention, in this post has been on the management transition at Apple, but management transitions are part and parcel of every company, with the changes sometimes forced on the company and sometimes voluntary. Expanding the discussion of management to the other companies in the Mag Seven can provide us with an opportunity to examine management transitions that have either already happened or that will happen in the future, and the ensuing frictions:

• At Microsoft, the only company in this group that traces its vintage back to Apple, there have been two CEO transitions, from Bill Gates to Steve Ballmer in 2000, and from Ballmer to Satya Nadella in 2014. While Gates built Office and Windows into cash cows, and Ballmer preserved them, Nadella created his own pathway to reincarnation by building up a cloud business that is now the dominant source of revenues for the company. By partnering early with OpenAI on LLMs, and investing massively in data centers, Nadella is now making a bet that AI can provide a further boost to the company’s operations, perhaps setting the stage for a second rebirth.

• Amazon has seen a management transition, where a legendary founder (Bezos) left the firm in 2021, and his successor (Andy Jassy) has taken the reins, with remarkably little fanfare. Like Nadella, though, Jassy is betting big on AI being a growth and value driver, and the success or failure of that bet will largely determine how his stint as CEO gets judged.

• Alphabet offers a case study of a company that tried to split the difference, by separating its cash cow (Google advertising) from its other businesses, naming Sundar Pichai as the CEO for Google, while remaining atop the other Google businesses (the bets in Alphabet). That experiment has struggled to deliver, as the other businesses remained earth-bound and in 2019, and Pichai took over as CEO of Alphabet as well. Over its lifetime, Alphabet has been immensely successful in coming up with products and services that catch public attention, whether it be its development of the Android operating system or its work on Waymo or Gemini, but it has struggled to convert those successes into revenues and operating profits.

• In three of the companies (Tesla, Meta and Nvidia), founders remain CEOs, though they bring very different perspectives and personalities into their roles. As I noted in my last post on SpaceX, Musk has veered between genius and eccentricity in his stewardship, but shareholders at Tesla have largely benefited from the rollercoaster ride. At Meta, Zuckerberg has been a shrewd businessperson in his management of his social media holdings, with savvy acquisitions of Instagram and Whatsapp boosting his ad-driven ecosystem, but he has also been headstrong in his pursuit of ventures that he feels are the “next big thing”. His expensive failed bet on the Metaverse led some investors to question his governance, and many of these investors worry that his bet on AI will play out similarly. Finally, on Nvidia, the company’s soaring market capitalization and huge success with AI chips has pushed Jensen Huang into the spotlight, but less than a decade ago, there were questions about his management as well.

The fact that all six of these companies have invested heavily in AI is a lead-in to what could be the key test for management at all of them. If the AI investments pay off and deliver value, Nadella will cement his legend status, Jassy will have created his own legacy at Amazon, the Alphabet experiment will finally pay off, and the founder-run companies will have more room to run. If the AI investments fail, though, Nadella’s reincarnation reputation will take a hit and Jassy’s position atop Amazon will be at-risk. The AI failure will also raise doubts about Alphabet’s capacity to grow beyond advertising and the rumblings about Zuckerberg’s big bets will get louder, but at these two companies, it is unclear what investors, no matter how large their holdings are, can do, since they have acquiesced to a voting share structure at these two companies that has reduced them to bystander status. At Alphabet, Brin and Page control 51% of the voting rights, with less than 10% ownership, and at Meta, Zuckerberg controls 57% of the voting rights, with about 13% of share ownership.

An Ode to Restraint

While there are many who compare to Tim Cook to Steve Jobs and find him wanting on vision and flair, I am grateful, as an investor in Apple, for the restraint and discipline that he brought to the job. That gratitude will stay intact even if Apple’s caution on AI turns out to be a mistake, since the restraint and rectitude that Cook brought to his job are management qualities that significantly undervalued. I don’t teach from or write cases, but I would love to see more business school cases about CEOs like Cook who are not easily swayed by the temptation of more growth and ego-driven acquisitions. I loved the Steve Jobs movie, but I don’t expect to see a Tim Cook movie anytime soon, and while that is understandable, it also explains why we will continue to have too many CEOs at companies viewing themselves as saviors, gambling shareholder money on turnarounds and rescues, when the better pathway would be acceptance and shrinkage. I believe that investors lose more money from companies trying to do too much rather than from them doing too little, and from overreaching than from underachieving.

YouTube Video

My posts on Apple

1. Apple: Thoughts on Bias, Value, Excess Cash & Dividends (March 1, 2012)

2. Apple: Know when to hold ‘em, know when to fold ‘em (April 3, 2012)

3. Emotions, Intrinsic value and Dividend Clienteles: The Apple postscript (April 6, 2012)

4. Apple’s Crown Jewel: Valuing the iPhone Franchise (August 29, 2012)

5. The Year in Review: Apple’s Universe (December 2012)

6. Are you a value investor? Take the Apple Test! (January 2013)

7. Back to Apple: Thoughts on value, price and the confidence gap (February 7, 2013)

8. Financial Alchemy: David Einhorn’s value play for Apple (February 8, 2013)

9. Apple: News, Noise and Value (April 30, 2013)

10. Love the company! Love the product! Love the stock! (September 9, 2013)

11. Watch the Gap: Apple’s Long and Twisted Journey (April 2014)

12. The Race to the Top: The Duel between Alphabet and Apple (February 2016)

13. Icahn exits, Buffett enters: Whither Apple (June 2016)

14. Apple: The Greatest Cash Machine in History (February 2017)

15. Investor Whiplash: Looking for Closure with Apple and Alphabet (December 2018)

My book on the corporate life cycle

1. The Corporate Life Cycle

The History of SpaceX

You may be surprised to hear that SpaceX is older than Tesla, at least in terms of chronological age, founded on March 14, 2002, in El Segundo, California. At its founding, Musk stated its goals as reducing the costs of space transportation and travel to Mars, but was viewed as having little chance of success by the space establishment, composed then of government agencies (NASA) and a few defense firms (Boeing and Northrop Grumman). SpaceX applied the lessons of modular engineering from the software business and it launched Falcon 1, its first space launch vehicle in September 2008; that successful launch led to a NASA contract for $1.6 billion, and rescued the company from near bankruptcy. In subsequent years, SpaceX developed Falcon 9, a reusable and heavier vehicle, with the Dragon Spacecraft unit, and became the first commercial entity to deliver cargo to the International Space Station. In 2013, SpaceX launched its first mission for a private customer, and quickly secured a dominant market share of commercial launch contract market. In recent years, SpaceX has invested in an even more ambitious version (in terms of size and power) of reusable spacecraft with Starship, and while its first launch in 2023 exploded in space, the company is clearly moving towards making it functional.

Along the way, the company added to its business mix, first with The Boring Company, a company that specialized in building tunnels that could be used to transport people, in 2017, before spinning it off as a separate entity. More significantly, in 2019, the company launched sixty Starlink satellites, with the end game of offering satellite-based internet services to customers, especially in areas where conventional internet service was limited. That endeavor has now grown to include thousands of satellites and had more than ten million active subscribers spread across the world, at the end of 2025. In February 2026, the company created its third business arm, with its acquisition of xAI, the parent to Grok, the Musk-developed competitor in the LLM space.

SpaceX had a slow start financially, as its initial years were spent developing the Falcon 1 rocket, and even after that development, the dependence on the US government and commercial satellite launchers resulted in revenues growing much more slowly than they did at Tesla, Musk’s other high-profile creation. Even as late as 2021, SpaceX reported revenues of just over $2 billion, almost entirely from its launch business, but Starlink’s subscriber based model has allowed revenues to increase more than five-fold since, reaching an estimated $15.6 billion in 2025, with just under 30% coming from the launch business ($4.1 billion) about the rest from Starlink subscriptions and related businesses ($11.4 billion); xAI, which was acquired in 2026, had subscription revenues of roughly $100 million in 2025. Without full financials to back up the statement, it is estimated that SpaceX generated an EBITDA of $8 billion in 2025, though with depreciation and other expenses considered, it is not clear how much (if any) operating (or net) profits the company delivered during the year.

On the funding front, Elon Musk used a portion of his winnings ($180 million) from his PayPal exit as seed money ($100 million) for founding SpaceX, but the company has required multiple rounds of venture capital to fund its infrastructure needs. The first venture capital round of about $12 million was in 2002, but there have at least thirty additional infusions amounting to more than $12 billion. While SpaceX counts big name venture capitalists in its investing roster (Founders Fund, Andreessen Horowitz and Sequoia), it has also seen increasing investments from public equity investors such as Fidelity and public tech companies such as Google. While these venture capital investments have diluted Musk’s ownership over the years, he continues to own about 42% of the equity in the company, and with differential voting rights, close to 80% of the total voting rights in the company.

Valuing SpaceX

It is true that intrinsic valuation, at least in its discounted cash flow avatar, is much easier to do at companies that have many years of historical data and peer groups of companies in the same business, and there are some who view one or both as pre-requisites. If you adopt that point of view, it is easy to see why so many view SpaceX as a company that cannot be valued (yet), since you don’t have access to even a single year of financials, let alone a long history, and there are no true competitors. In fact, it is likely that even if the financial statements are made public in a prospectus, most will continue to avoid valuing the company, using uncertainty about the future as an excuse. If you define intrinsic value as the value of a business based upon its capacity to generate cash flows in the future, there is nothing in that definition that requires either historical data or peer group information, and statistically, the fact that you face uncertainty or that you are missing information does not imply that you cannot make estimates, just that the estimates will be noisier.. I have long argued that you can estimate the value of young companies with minimal data, as long as you build a valuation around a business narrative, and accept that this valuation will change, as circumstances do, and with SpaceX, I will get a chance to put this argument into practice.

To value SpaceX, I consider each of its three core businesses separately since they differ not just on operating metrics, but also on the competition faced in each one.

• The launch business, which is where SpaceX was born, is still its most identifiable business, and frames the company’s story not just as a futuristic company, but one that was able to overcome some of the most significant technological challenges of any start up and not just survive but thrive. SpaceX has established such a robust and long-standing cost advantage over its competitors in the business, stemming from its existing infrastructure investments and reusable rocket technology, that it had a market share in excess of 80% of the launch market in 2025. Its competition will come from some private and government-funded players, who may be able to capture market share, notwithstanding their higher costs, due to security and nationalistic concerns..

  • Valuation narrative: The space launch market is estimated to be about $30 billion in 2026 and is expected to grow to $100 billion in 2036, as government and private business demand increases. SpaceX will continue to dominate the business, albeit with a slightly less dominant market share (70%, down from >80% in 2025) of the total market and as costs decrease with scale, operating margins will increase over time to 40%. (I am being conservative in my estimates, insofar as I am ignoring space travel and expanded business opportunities in space, but I don’t think, at the moment, that either offers a viable path to augmenting revenues).

• The internet service business, built around Starlink, is the business that accounts for almost two thirds of the revenues of the company in 2025, and it builds on the infrastructure built for the launch business, since SpaceX has used it to launch thousands of satellites into space. At the end of 2025, Starlink had close to 10,000 satellites in space, about two thirds of the entire global count, and is adding to that number every month. That has allowed it to double its subscriber numbers to just over 10 million, in the last year, and while Amazon’s acquisition of GlobalStar has brought a potential competitor into the mix, GlobalStar has a fraction of the satellites that Starlink does. The challenge for any satellite-based internet service provider is that notwithstanding the use of low-earth orbit (LEO) satellites to improve service, the broadband service lags more conventional internet technology (fiber optic and cable) in much of the world, leaving it (at least for the moment) with a niche market of rural areas, countries with damaged or no infrastructure and people on the go (airplanes, trains and cars). Thus, while the total internet service market is estimated to be close to a trillion and a half dollars globally, in 2025, satellite-based service accounted for about 1% of that market, delivering under $15 billion in revenues, with StarLink having a dominant market share.

  • Valuation narrative: The satellite internet services market will continue to grow, as technology improves service quality and transit demand for better wifi grows, from $15 billion to $160 billion (from less than 1% to 10% of of the internet service market) over the next decade. Starlink will see more competition, but its lead in satellites and capacity to use its launch business to get more into space, will give it a substantial advantage and a market share of 75% of the overall market). The cost of customer acquisition will ease over time, as business customers become a larger portion of the business, and operating margins will approach 60% in steady state, as the unit economics are very positive.

• The Large Language Model (LLM) business, from the acquisition of xAI, has brought AI into the SpaceX story, and while that may add to the pricing excitement, Grok lags the other LLMs in terms of revenues and reach, for the moment. In terms of usage, Anthropic (with Claude), OpenAI (with ChatGPT) and Google (with Gemini) are not only more widely used than Grok, in business setting, but are further along in converting them into revenues. While Grok has been bundled into the X Premium subscriptions, and earned about $80 million in revenues in 2025, it seems to be focusing more on consumers and only on niche portions of the business market. This is the most diffuse and volatile of the three markets, in terms of potential market, since the potential market can run from the tens of billions (if they remain subscription-based) to hundreds of billions or even trillions (if they become replacements for human labor or massive productivity boosters). It is possible that Grok may concede the larger and more competitive business space to Claude and ChatGPT and focus instead on consumer subscriptions, giving it a smaller market, but one with less competition.

  • Valuation narrative: The overall LLM market will continue to grow, but more in business applications than in consumer apps, with the market size being determined by how well AI can replicate human labor and regulatory restrictions. xAI will target primarily consumer subscription revenues and niche business applications. That will give it smaller revenues ($80 billion in 2036) than its LLM competitors, but one with less competition and higher margins (50% operating margin), and less reinvestment as it avoids going head-to-head with Anthropic, OpenAI and Google for business use.

• It is undeniable that SpaceX, as a lead player in three fast-growing and volatile businesses, may be able to use its infrastructure to expand each of these businesses. The space launch business, which has generally focused on delivering commercial or government loads and satellites into outer space may become a springboard for space travel, for leisure, research or business. With its satellite broadband offerings, the possibility exists that the technology and the reach will improve to a point where the service can compete with fiber-optic and cable broadband offerings, perhaps at much lower cost. With xAI, the possibility that Grok finds a way to outflank its LLM rivals, including Claude, Gemini and ChatGPT, in terms of business offerings may be low, but it does exist. The recent acquisition of Cursor, a young AI company in the coding space, suggests that xAI has not thrown in the towel on business applications. In truth, these are all options that may not be viable at the moment, but if they become viable, could add immense value.

  • Valuation narrative: This part of the story is built on the expansion options that SpaceX has to enter large markets, with low probability and high payoff. In a crude attempt to capture this part of the story, I will attach an expected revenue in 2036 to these other businesses of $50 billion and an operating margin of 30%. Since these expansion options, if they do exist, will not show up in the near future, the revenues from these options ramp up after year 6 (2032) in the valuation.

Pulling these storylines together as valuation inputs, I estimate the following numbers for 2036, for the three business lines and for the expansion options category:

With these numbers in place, and using a cost of capital reflective of SpaceX’s business mix (of aerospace/defense, telecom services and AI) of 8%, we can estimate the company’s value:

Download spreadsheet

With my story and inputs, the value that I derive is $1.22 trillion, about 10% below the private market pricing and about a third below the expected IPO pricing, but still astonishingly high for a company with $15.5 billion in revenues in the most recent year, and a host of question marks about corporate governance. Note that in this iteration, I have ignored cash and debt, since I do not have the company’s financial statements, but it is unlikely that either will have much of an impact on the value of equity for a company with this high a value for its operating assets.

As you can see, the SpaceX story not only has many moving parts, but is fraught with uncertainty, and without full financials, it does not have a good starting point. That said, as the story plays out, we will get more clarity, and the story will need to be reworked, with the value consequences unclear. For the moment, though, I am uncertain about every input in my SpaceX valuation, but uncertainty is a continuum, and I am less uncertain about some inputs (such as the revenues and margins in the space launch and satellite internet service businesses) than about other inputs (including the revenues and margins of the LLM and expansion businesses). If you are wondering why the cost of capital is only 8% for the company, close to the median cost of capital for a US company, it is because much of the risk here is specific to the company (thus reducing the effect in a diversified portfolio) and cuts in both directions (upside and downside). In fact, if there are outliers, they are more likely to be on the upside than the downside. With these considerations in mind, I tried to be open about how uncertain I feel about my estimates, and the results of a simulation yields the following distribution for value:

Since the simulation is centered on the same expected values for inputs as I used in my base case, it should come as no surprise that the median value, across ten thousand simulations, of $1.29 trillion is close to the base case valuation of $1.22 trillion. As the pricing for the IPO starts to gain traction, it is worth recognizing that a $1.75 trillion or even a $2 trillion pricing falls in the range of the distribution, though with little or no upside left for an investor paying that price.

If you feel that it is best to wait for the prospectus to be filed, before doing the valuation, I understand but there are three points worth remembering. First tt is unlikely that the prospectus will contain data that will move the intrinsic value story, since none of the numbers in the reported statements will be large enough to alter the immense value coming from expectations of future growth. Second, while the prospectus will contain estimates of total addressable market and perhaps even profitability, in my experience, it will be hype; expect to see trillions of dollars thrown around nonchalantly for market size. Third, as I see it, it is not an either/or proposition, since I can value the company now, and revisit the valuation when the prospectus comes out, with the advantage being that you are less likely to be swayed by the sales pitch in the prospectus.

Pricing SpaceX

In an initial public offering, companies are priced, not valued, by bankers (for the offering price), by investors (as the stock starts trading) and by observers to make judgments (on whether it over or underpriced). Thus, you can make the argument that the valuation, with all of its moving parts, is irrelevant, and that you should price SpaceX, not value it, if your intent is to trade on the IPO. That is a legitimate critique, but the argument that pricing somehow dispenses with the need to make assumptions about market size and profitability or does not have the same uncertainties is not. You can take issue with the intrinsic valuation because of the layers of assumptions that I had to make along the way, and I know that for many investors, either invested already in the company ,or planning to invest in it, a pricing may seem less daunting. While I sympathize, I am afraid that the uncertainty will be just as much of an issue in pricing SpaceX, and to see why, take a look at the steps in the pricing process:

At each step in the process, you will run into issues.

• On the pricing metric front, the problem with picking a metric is an information vacuum, with only two scalars, revenues and EBITDA, available, and even if you consider the most recent private market pricing, which priced the company at about $1.25 trillion, as the market value of equity, the absence of debt and cash numbers makes it impossible to back into enterprise value. This problem should be resolved, for the most part, when the company files a full prospectus, but for the moment, if you assume that the net debt number is close to zero, the resulting enterprise value of $1.25 trillion yields nosebleed multiples of 81 times revenues and 156 times EBITDA for the company, using 2025 numbers. Cven with minimalist information, there will be pricing variants that use expected revenues (or EBITDA) in a future year as a scalar, as can be seen in this graph:

  

There is no inherent problem with using forward numbers in pricing, as long as you do the same for all of the companies in your peer group, but in the case of SpaceX. it is inevitable that bullish analysts, unable to justify the sky high pricing values with trailing 12-month numbers, will resort to using forward pricing, and add to the bias, by inflating revenues in future years. In the graph, I have used the forecasted revenues in 2030 and 2035, from my intrinsic valuation, and the EV to Sales ratio drops from 80.13 (112.18) to 3.91 (5.47), using the private company (estimatedIPO offering) pricing of $1.25 trillion ($1.75 trillion) as the enterprise value.

• The even bigger challenge in pricing SpaceX will be in the second step, where you have to find comparable firms (or a peer group) to base your pricing on. There is obviously no company out there that is remotely similar to SpaceX, as a composite company, and even if you break it down into businesses, it is likely that you will hit roadblocks. On the space launch business, using publicly traded aerospace and defense companies like Boeing and Northrop Grumman is a non-starter, because they are low growth, lower-margin businesses, unlike SpaceX. Palantir may seem like a logical alternative, and while it may have high growth potential, it is a data/software company that does not have the infrastructure needs that SpaceX does. On the internet service business again, there are conventional telecom firms like Verizon and T-Mobile, but the economics of their offerings are different, and they are not growth companies. In April 2026, I computed the multiples of revenues and EBITDA that publicly traded companies in the aerospace/defense, internet services businesses and technology companies trade at, and they are far lower than what SpaceX can be expected to trade at, if it goes public at $1.5 to $2 trillion:

  

• Finally, on xAI, there are other LLMs, almost all of which are private companies now, and while you can use the pricing from most recent VC rounds, you are on shaky grounds. Even if you forged ahead with a peer group of high-growth, tech companies, your pricing will almost certainly have to revolved around revenues, rather than profits, and based upon very small samples.

• The one part of pricing that you will almost certainly see is the story telling, especially with analysts who are locked into finding SpaceX to be a buy, with the pricing more of an ex-post rationalization than a analytical tool. No matter what peer group you pick, SpaceX will be priced higher than comparable firms, and to back the argument that it is still a good investment, you will hear stories of its large potential market, significant competitive advantages and profitability. All of these stories are grounded in truth, but they are empty if they remain stories. I will predict that there will be far more buy than sell or hold recommendations for SpaceX, when it does go public, with analysts doing pricing gymnastics with forward multiples, hand-picked peer groups and fairy tales to justifying their positions.

Pricing is an exercise in data analysis, and any pricing of SpaceX will reflect the statistical limitations of the data. Put simply, the only difference between intrinsic valuation and pricing, when it comes to the uncertainties you face with SpaceX, is that with the former (intrinsic valuation), you have to face up to the uncertainties and make your best explicit judgments (on revenues, margins and reinvestment), whereas with the latter (pricing), they remain implicit. If bias is your biggest adversary in assessing SpaceX, and you use pricing, it is very likely that you will find a pricing metric and hand-picked peer group to reflect your biases, and then tell yourself a story on why SpaceX is cheap or expensive.

The Bottom Line

SpaceX is an engineering marvel that has shown its naysayers, which included almost every luminary in the space community, to be wrong. That said, for potential investors, there are lessons to be learned from watching Musk’s stewardship of Tesla. As with all of Musk’s creations, SpaceX will be a shape shifting entity, frustrating investors who expect companies to follow linear paths in the corporate life cycle, going from young growth to maturity; it will shift from one narrative to another, often with no advance warning, causing whiplash for investors. When I bought Tesla in 2019, after its stock had taken a beating, I described the company as my corporate teenager, and with Musk in full control, SpaceX is likely to follow the same unpredictable path. That makes it a difficult company to buy, but it makes it an even more dangerous company to sell short, as Tesla short sellers have discovered in the last two decades. With all that said, SpaceX is a unique company with immense competitive advantages, and while I would not be interested in buying at the rumored IPO pricing of $1.75 trillion, it is one big correction away from being fairly priced or even cheap. If that happens, I will be a buyer, but will do so with the recognition that this company comes packaged with a founder who is both uniquely gifted and deeply flawed, and complaining about the parts of Musk you do not like, while enjoying the fruits of the aspects that you do, is unfair.

YouTube Video

Spreadsheets

1. Valuation of SpaceX in April 2026 (Pre-IPO, and with limited financials)

The Market Narrative in March

We live in an age of commentary, as self-proclaimed experts offer prognostications, half-baked or otherwise, about what is to come, and the Iran war, with its mix of politics, economics and religion baked in,, has drawn a large and extremely diverse set of expert forecasts. Given the strong priors (about Iran and Trump) that many of these experts bring to the game, it should not be surprising that their views about how the war will play out and the effect on markets is driven by those priors. It is up to markets to reconcile these contradictory perspectives, and come to consensus, and I will try to extract from market behavior what the market narrative is, leading into April 2026, with the recognition that it could be wrong and change overnight in good and bad ways. That said, over the last decade, I have learned that the market is far better at making sense of complexity and uncertainty than experts are, and it behooves us therefore to listen to what it is saying.

The Oil Price Shock

As with almost every event in the middle east, the effects of the Iran War played out first in oil prices, and oil has been the lead player in March, surging and volatile, but with disparate impacts even within that market. In the graph below, I look at spot prices on Brent Crude and West Texas Intermediate (WTI) during March:

Both Brent and WTI crude oil saw prices increase in March, but with the price of Brent rising 49.9% and WTI rising 48.6% during March, the difference between the two almost doubled during the second half of the month. That divergence reflects the two-fold effect of the war on oil supply, with the first being the shuttering of oil production in the Gulf States and the second being the effecting throttling of ship traffic through the Strait of Hormuz, a key passageway for Middle Eastern oil to Asia and Europe. While both factors push up oil prices, oil and gas production in the US, the largest oil producer in 2025 (producing 13.58 million barrels or 16% of the total), was less affected by the Hormuz closing and supply chain issues, explaining the increasing price divergence mid-month.

There was another tea leaf to read, and it came from watching oil futures prices. In the graph below, I compare the spot prices to Brent crude to June and December futures contracts prices:

While spot and futures prices have both risen in March, the latter have gone up less, indicating that, at least for the moment, the market sees the interruptions in oil supply as more temporary than permanent, though the market does see a lasting impact even in that optimistic scenario, with December futures up almost 25% over the pre-war level.

Inflation, Interest Rates and the Economy

The creation of OPEC and the oil price embargo in the 1970s and the subsequent inflation spiral in the 1970s is now part of market legend, and the interlude in 2022, when the Russian invasion of Ukraine, and the subsequent sanctioning of Russian oil, caused a spike in inflation rates, has made investors wary. While the effects on gasoline prices are in the news, it is one item in the inflation basket, and it is unclear still how much higher oil prices will affect inflation for the rest of the year and perhaps into next year. While we wait for the actual inflation numbers to come out, markets don’t have that luxury and the early and perhaps best indicator of market expectations on inflation are showing up in interest rates. The graph below looks at 3-month and 10-year US treasuries over the course of March 2026:

The 3-month treasury bill rate has barely budged over the month, moving from 3.67% on February 27, 2026 to 3.70% on March 31, 2026, but the ten-year bond rate saw a much bigger increase from 3.97% on February 27, 2026, to 4.30% on March 31, 2026. The biggest increases in rates are in the intermediate maturities, with the 2-year and 5-year rates rising by 0.41% over the course of the month. If you view interest rates, as I do, as driven by expected inflation and expected real growth, the most plausible reading is that the market sees an increase in inflation that is persistent. If you are a Fed-watcher, though, your reading may be that the rise in oil prices has tied the hands of the Fed, lowering the likelihood that the Fed Funds rate will be cut in the coming months, but that would leave you with a puzzle to resolve. Since the Fed Funds rate, an overnight bank borrowing rate, has its biggest impact on the short end of the maturity spectrum, how do you explain the fact that short term rates have not changed much?

The increase in interest rates is not just specific to the US, with rises in rates across other currencies, as you can see in this graph of ten-year Euro, Yen and Yuan rates:

The Japanese Yen and Euro rates are up significantly over the month, but the Yuan rate has seen no change in March 2026. Staying with the market narrative, this indicates higher inflation across countries and currencies.

While there are many who are speculating on what higher inflation and oil prices will do to the economy, and investment banks and data services (See Moody’s, Goldman Sachs) have been rushing to update their forecasts for the US economy, the market has not been in as much of a rush to make the judgment. The economy was showing signs of fatigue coming into March 2026, with anemic growth and employment numbers, and it is possible that the oil price shock will tip it over into a recession.

The Price of Risk

The heightened uncertainty generated by war and its consequences has played its way out not just in oil prices and treasury rates, but in the prices that investors charge for risk. In a month where the clash between greed and risk took front stage, with the balance shifting often on a minute-by-minute basis during the trading day, we also see increases in the price that investors charge for taking risk in both equity and bond markets. In the equity market, that price of risk is the equity risk premium, a topic that I talked about extensively in this post and paper, with the argument that a good measure of this risk premium will be forward-looking and dynamic. My implied equity risk premium estimates tried to capture the changes in equity risk premiums on a daily basis, and completing the assessments for the entire month, here is what the equity risk premiums looked like in March 2026:

The surprise here is not that the equity risk premium rose over the course of the month, expected given what was happening in the Middle East, but that it rose so modestly. In fact, over the course of March, the implied equity risk premium for the S&P 500 rose from 4.37% on February 27, 2026, to 4.77% at close of trading on March 31, 2026, an increase of 0.40% for the month.

In the bond market, the price of risk is the bond default spread, and in the graph below, I look at default spreads for seven bond ratings classes from AAA to C (& below):

Here again, the spreads increased over the month, but only modestly, even at the lowest ratings classes. Thus, the BBB default spread over the 10-year treasury rose only 0.08% during the month, from 1.07% on February 27, 2026, to 1.15%on March 31, 2026, and the high yield spread (for CCC and below) increased from 9.50% at the start of March 2026 to 10.10% at the end of the month.

The third proxy for risk is the volatility index (the VIX) for US equities, and that measure rose during the course of March 2026:

During March 2026, the VIX rose from19.86 at the start of the month to 25.25 by the end of the month, an increase much smaller than the increases we saw in March 2020 (COVID) or in the first week of April 2025 (Tariff week).

With the caveat that this is still mid-narrative, the bottom line from the movement in all of these risk measures is that while the market had a bad month, much of the marking down in equity values can be attributed to real concerns about higher inflation and economic damage, and is not the result of panic selling, at least in the aggregate. To back this up, I took a look at two collectibles - gold, which has a history of holding its value or even increasing during crises and panics in financial markets, and bitcoin, which has not had that history so far, but is marketed by its advocates as a potential hedge:

Gold was down 10.42% during March 2026, uncommon for a crisis month, but bitcoin was up 3.30% during the month, and it is entirely in keeping with bitcoin investors marching to their own music, though it will be interesting to see how this dynamic plays out, as this repricing continues.

Effect across Geographies

The war is in the Middle East, but there is no place to hide from its effects. To see how the war has played out in different regions, I looked at the change in aggregate market cap, in US dollar terms, in March 2026:

You may be surprised to see Africa & the Middle East and Eastern Europe & Russia show up as the best performing markets, with about 2% decreases in market capitalization, but it reflects the dual impact of the war. While it has wreaked havoc across the Middle East, the higher oil prices that it has brought with it are providing upside for oil producers that offsets some of the damage.

Since these dollar returns reflect local market performance as well as the strength/weaknesses of their currencies against the dollar, I looked at the US dollar’s performance in March 2026:

I know that I am piling on at this stage, but I do compute equity risk premiums for other countries twice a year, once at the start and once mid-year. Given how much March has shaken up the status quo, I will make an exception and re-estimate equity risk premiums, by country, updating both my mature market premium (which I estimate from the S&P 500) as well as the country ratings, default spreads and country equity risk premiums for other countries.

Download spreadsheet with country ERPs

It is worth noting that these equity risk premiums are computed based upon sovereign ratings, which are slow to change, as the world convulses. That has been an issue with my ERP computations for Russia and Ukraine, since 2022, with the rating for the former withdrawn and the rating for the latter frozen at Ca (Moody’s); I have use a country risk score from PRS for the last two years to update Russia’s equity risk premium, an have done the same for the Ukraine in this update. You can see the same issues now, with the war in Iran rocking the boat, and at least for the Middle East, there is reason to believe that the ratings may understate country risk. While none of the countries in the war zone have seen their sovereign rating change (yet), these countries have market estimates of sovereign default risk in the form of sovereign CDS spreads, I looked at the movement in those spreads during the course of the month:

Not surprisingly, market measures of default risk are more sensitive to war effects, and have risen for much of the Middle East, as worries have mounted, with bigger increases in Qatar, the UAE and Turkey than in Saudi Arabia and Kuwait. The United States has also seen a surge in its sovereign CDS spread, and the global sovereign CDS spreads have risen about 12% in the first quarter of 2026. Using these sovereign CDS spreads as measures of default spreads for this part of the world may yield more realistic equity risk premiums.

What now?

I noted at the start of this post that the uncertainties that manifested during March 2026 about the direction, duration and effects of war are still unresolved and perhaps even grown as we start April. As investors try to navigate their way through this period, here are the questions that you will need to answer to decide where you fall in the continuum between complacency to full-blown panic:

In the complacency scenario, the war ends quickly (in days or weeks, rather than months), the damaged infrastructure is repaired quickly and the new regime in Iran is viewed favorably by the rest of the world, allowing the sanctions on the country to be removed, it is likely that oil prices will drop, perhaps even to below pre-war levels, as Russian and Iranian oil is freely bought and sold. In the full-scale panic scenario, the war continues for months, with lasting damage to infrastructure and supply chains and Iran’s new government stays sanctioned, oil prices are likely to stay high and perhaps even go higher, the global economy will be kneecapped and parts of the Middle East (Dubai and Abu Dhabi) that had created a business and tourist friendly setting will struggle to find their balance.

In either case, the war has shaken up the status quo, and I see lasting consequences that go well beyond oil. The capital flows from the oil rich countries which has flowed generously to everything from AI start ups to Premier League clubs will shrink, creating down-market effects. That money, and the funds that were set aside to build vanity projects, from ski resorts in the deserts to state-of-the-art cities will be redirected to building pipelines and securing the flow of oil. Global politics has also been roiled, and even if the war ends quickly, there is damage that has been done to partnerships and security agreements that cannot be undone.

YouTube Video

Datasets

1. Equity risk premium for S&P 500, by day (March 2026)

2. Equity risk premiums, by country (updated April 1, 2026)

The Origins

In the late 1990s, I was approached by the Stern School of Business with a request to serve as the organizer for a class on investing, where MBA students would spend a session a week, for a semester, hearing from successful investors of all stripes, and discuss what they learned from that talk in a second session each week. Over the course of the semester, the class had fourteen speakers, and because of our New York location, it drew from a range of investing types. Thus, students heard from a well-known value investor one week, the manager of one of the best-regarded growth mutual funds the next, a high-profile technical analyst in the third, and so on. The speakers approached investing in very different ways and had different perspectives on financial markets and how to exploit market mistakes, but they all had been successful as investors.

As I led the discussion of each speaker’s market views and investment practices each week, I noticed students in my class developing whiplash, as they instinctively try to incorporate the views and practices of each speaker into their thinking. As the weeks went on, that became a problem, since other than investment success, the speakers shared little in common, and their views about markets were sometimes contradictory. By the end of the class, there was a fairly large subset of the students who ended up more confused by what they had heard during the semester, rather than enlightened. As I reviewed the class, before handing it off to someone else, I took an inventory of what I had seen not just in the class, but in investing in general, and came to the following general judgments about investing:

1. There are very few active investors, who win consistently over time: Active investing is one of the most difficult games to win at, and one reason is that you match the average investor, effortlessly and almost costlessly, by investing in index funds. Active investing has the unenviable task of trying to be better than average, and by enough to cover the costs (research, data, personnel, transactions) associated with being active. Just as illustration of how much of a mountain this is to climb, take a look at the percentage of active institutional investors who beat their respective indices over the last decade

   

   :

   While there some active money managers who “beat the market” over a year, two years or even five, very few are able to hold on to these excess returns as you lengthen their active investing stint. Like gamblers in a casino, who strike it lucky early, but stay gambing too long, they often leave with none of their gains, or worse. Before I get a blowback, I am fully aware that that there are investing legends (Warren Buffett, Jim Simon and George Soros, to name just three), but the very fact that we can name them suggests that they are the exceptions, not the rule.

2. Even with those successful few, it is very difficult to separate luck from skill: Much as investment books and classes claim otherwise, investing results are affected by so many forces that are out of your control that disentangling how much of your final returns can be attributed to skill and how much to luck is very difficult to do.

3. These successful investors have widely different pathways to delivering success: If you were to make a list of the investors who have had the most success in markets in the last century, I would wager that you would be looking at a very diverse group, not just in terms of how they succeeded, but also in terms of personality. The three investors I named as legends - Buffett, Simon and Soros - obviously had very different views on markets, and how to exploit market mistakes, but even with investors who are often viewed as being from the same grouping, differences remain. Buffett may have learned his early lessons from Ben Graham, but the Graham and Buffett approaches to value investing are varied, with the former more focused on screening for cheap stocks and the latter more interested in finding companies with solid moats and great management.

4. Imitating successful investors does not seem to provide much payoff: The practices of successful investors have been probed and investigated by other investors and journalists, and some of them have dozens of books that claim to tell you the secret of their success. Warren Buffett is perhaps the winner in this race, with not only a multitude of books that track his investing life but also his annual letters to Berkshire shareholders which laid out his investing perspective in detail. That said, the investors who tried to follow in his footsteps, often imitating every aspect of his approach, have, for the most part, not been able to match his success.

My takeaways from these assessments are two fold. The first is that there can be no one dominant investment philosophy that is the best for all investors, and any claims to the contrary, whether it be for value investing or market timing or trading, are disingenuous. The second is that there is a right investment philosophy for each individual that reflects that individual’s views and beliefs about markets and characteristics as a person.

The Core Idea

The recognition that each investor needs an investment philosophy that is tailor-made to his or her beliefs and personality became the starting point for my creating a class, and writing a book, about the topic. Before I describe what I try to do in the book, I should start with a definition of what I mean by an investment philosophy, and perhaps the best way to do that is by describing what it is not. First, an investment philosophy is much richer and more complete than an investment strategy, with the latter often coming out of the former. Thus, applying a screen to find stocks that trade at low multiples of earnings (low PE ratios or low multiple of EBITDA) is an investment strategy, but the investment philosophy that gives rise to that strategy is one that is built on markets under pricing companies with low growth or boring businesses, perhaps because investors are dazzled by growth and drawn to the excitement of newer businesses. Second, an investment philosophy is not an investment slogan. “Buy low, sell high” is an investment slogan, and a meaningless one at that, since that is the end game of almost every investment philosophy.

If you have been investing for a while, and have never stopped and asked yourself what your investment philosophy is, it is understandable. In fact, you may wonder why you should constrain yourself to an investment philosophy instead of looking for bargains wherever you can find them. The problem with not having a core philosophy is that is exposes you, as an investor, to a whole host of consequences, most of which are negative:

1. Chasing winners: If you don’t have an investment philosophy, it is almost a given that you will find yourself drawn to whatever strategies worked best in the recent past. Your portfolio will suffer from whiplash as you chase last year’s winners, whether that be the Mag Seven or technology stocks or small cap stocks, and while your turnover and transactions costs rise, you will have little to show in terms of returns.

2. Scam target: Greed is universal, and that leads us to look for ways to make lots of money with very little risk. Without an investment philosophy constraining you, you will be an easy mark for investment scams, drawn in with promises of upside with little or no downside.

3. Empty investing cupboards: If you do find an investment strategy that works at delivering returns, it is worth remembering that the clock is ticking, and that imitation and market corrections will cause that strategy to stop working, sooner rather than later. If that is all you brought to the market, your investing cupboard will be empty and you will find yourself running to stay in place. The advantage of having a coherent, well thought through investment philosophy is that you can go back to it and mine it for other strategies that may exploit the same market mistakes. Thus, if your investment philosophy is that markets undervalue boring, low-growth companies, and low PE ratios are no longer doing the trick (of finding cheap stocks), you may look for other screens (low volatility) that find you boring companies that are mispriced.

Simply put, every investor needs an investment philosophy to guide him or her in the difficult task of trying to delivering success.

Rather than create a laundry list of philosophies, I will use the investment process as the vehicle to describe how and where the different investment philosophies emerge from, as well as diverge:

Using this process, the choices in investment philosophies emerge:

1. Active investing versus Passive indexing: If, as we noted in the last section, doing nothing can deliver returns approximating the average, and nine out of ten investors who try to beat the average fail, there is no shame in adopting a passive indexing philosophy, where your allocation across asset classes is determined by your risk aversion and need for liquidity, and index funds fill out the rest of the dance card. It is human nature, though, to seek to be better than average, and it is perhaps that desire that drives many into active investing choices, and there are multiple pathways that they can adopt.

2. Investing versus Trading: The second divide in investing philosophies comes from the difference between value, which is driven by cashflows, growth and risk, and price, determined by demand and supply. Investing requires assessing the value of an asset, buying if the price is lower than that value and selling if it is higher, and waiting for the gap to close. Trading, on the other hand, is about gauging market mood and momentum, buying if you expect those forces to drive the price up and selling otherwise.

Within each of these groupings (investing or trading), there are sub-groupings. Trading can take different tacks, depending on where you think that market mistakes lie. The first, price traders, use the information on prices and trading volume to detect shifts in mood and momentum, with charts and technical indicators as tools, to try and generate profits. The second group, information traders, trades around information releases, such as earnings reports, acquisition announcements or even insider trades, with some trading ahead of the news, some at the time the news is announced and some in the aftermath, all trying to take advantage of what they see as market mistakes in reacting to that information. The third group, arbitrageurs, focused on finding the same or related assets trading on different markets, looking for mispricing across these markets, and locking in that mispricing as excess returns.

Investors, for instance, can be drawn to value or growth, and while that difference is often stated in terms of pricing multiples, with value investors buying low priced stocks (low PE, low price to book etc) and growth investors drawn to higher growth and high priced companies, I prefer to think of the differences in terms of where each group thinks it can find bargains. Using my financial balance sheet construct, where I divide the value of a firm into the value of investments already made (assets-in-place) and investments anticipated in the future (growth assets), value investors view their odds of finding market mistakes to be greater with assets-in-place, whereas growth investors feel that their odds are better in finding misvalued growth assets:

Within value and growth investing, there are further sub-divides. Value investing can span the spectrum from passive screening, where you screen for stocks that have specific characteristics (low PE, high growth, high ROE) and label them as cheap, to more activist poses, where investors with deep pockets (individual activist, private equity funds) not only take positions in companies that they believe are under or over valued, but also push for change at these companies. Growth investing has its own version of activist investing, in the form of venture capital, invested in young, growth companies, where in addition to supplying capital for growth, venture capitalists take an active role in how these companies evolve over time and exit the marketplace (IPOs, sale to another company).

3. Market Timing vs Stock/Asset Picking: In market timing, your focus is less on individual stocks or assets and more on deciding whether a market (equities, bonds, real estate etc.) is under or over priced. Returning to the investment process, your focus is on allocating your portfolio across asset classes, based on your market views, underweighting “expensive” asset classes and overweighting “cheap” ones. In stock/asset picking, you take the market as a given and try to find the best individual investments within each investment class for you - the cheapest stocks, bonds and real estate that you can find. There is an ironic contradiction in making this choice. It is undeniable that a successful market timer will make far more money than a good stock picker, but it is also true that it is much more difficult to be a successful market timer than it is to be a good stock picker. The picture below captures the choices in terms of investment philosophy, framed in terms of where they enter the investment process:

Even if you feel that you have an investment philosophy in place, I think being aware of how others approach markets and keeping an open mind, where you borrow parts of other philosophies and incorporate them into yours will make you a better investor.

Finding an Investment Philosophy

Looking at the menu of investment philosophies, from passive indexing to arbitrage, my end game in my book and for the class on investment philosophies was not to advance a single philosophy or even compare them, but to provide as unbiased and complete a picture, as I could, of the data backing each philosophy and more importantly, the personal characteristics that you would need to succeed with that philosophy.

Step 1: Views on Market Mistakes and Corrections

The first step in finding your investment philosophy is with a view of where (and why) markets make mistakes, and how they correct them. Even the firmest believer in efficient markets will concede that markets not only make mistakes, but sometimes make big ones, but the divergence between them and active investors lies in the nature of these mistakes. In an efficient market, market mistakes will be random, and since there is no systematic pattern to them, there is no pathway for active investors to find these mistakes, even with access to data and powerful tools. Active investors, in contrast, believe that there are systematic patterns that you can use to find these mistakes, and to exploit them for profits, with traders believing that those patterns are in the pricing and volume data and investors hewing more to fundamentals. That said, active investors can and will disagree about the types of market mistakes, with some buying into the notion that markets learn slowly, whereas others believe that markets overreact, and it is healthy for investors to have these disagreements.

Step 2: Pick an investment philosophy that reflects market views

Your views on market mistakes and corrections should guide you in your choice of investment philosophies. Thus, if you believe that markets overreact to news, good or bad, you may decide to become a contrarian, either trading (by buying after bad news and selling after good) or by investing (by buying companies with solid fundamentals whose stock prices have dropped by far more than they should have). Conversely, if you believe that it is momentum, not fundamentals, that is the biggest drivers of stock price movements, you may choose to ride that wave, based on charts and technical indicators. Superimposing time horizon onto the types of mistakes that markets make, you can create a matrix of investment philosophies:

Do you have to pick a single philosophy? Not necessarily! You can meld two or even more than two philosophies together, as long as you meet two conditions. The first is that the melded philosophies have to share a core belief about market mistakes. Thus, if you believe that market s overreact, you can be a contrarian value investor, buying companies that have been beaten up in markets but have intact fundamentals, and timing your purchases right after bad news releases, when markets overreact. The second is that you have to identify which of the philosophies is your dominant one, and which one is secondary, allowing you break ties where the two push you in different directions. Staying with the melded contrarian philosophy, and assuming that the contrarian value philosophy is your dominant one, you will choose to not to buy a stock that is down 15% after a bad earnings report, if it is still trading closer to its highs than lows.

Step 3: Check for viable strategies

Investment philosophies are a critical component, but to make money on a philosophy, no matter how well thought through, you need to devise investment strategies that can generate profits for you. In coming up with these strategies, you will confront the two realities that cause many strategies that look good on paper to fail: transactions costs and taxes.

• On the transactions cost front, the brokerage trading cost is just a small part of the overall cost, with two other costs that can often be much larger. The first is the bid-ask spread, small for large, very liquid stocks, but much larger for smaller and less liquid investments. The second is price impact, again non-existent if you are a small investor buying or selling shares in a large market-cap company, but substantial if you are a large investor trading on an obscure stock.

• On the tax front, some strategies will create more tax costs than others, partly because of how investment income is taxed (dividends create immediate tax consequences but capital gains require trading to incur tax liabilities) and partly because of how much trading your strategy will require of you, with higher turnover generally creating more tax liability.

If you are planning on being an active investor, there is one final skill set that you will need to acquire, and that is the capacity to test whether a strategy can beat the market. The volatility in returns can sometimes create illusions, where a strategy looks like it is delivering excess returns, but those returns are almost entirely due to statistical noise.

Step 4: Check for personal fit

Investment philosophies, and the strategies that emanate from them, come with different demands in terms of time horizon, with some requiring holding on to investments for many years and others requiring trading in minutes, different risk exposure and divergent tax consequences. Investors who choose to adopt these philosophies have to reflect on whether they are good matches, on the following fronts:

1. Capital to invest: If you are just starting on your investment journey, and have only a small amount of capital to invest, your choices in terms of investment strategies narrow. You will definitely not be able to be an activist investor, since you will have no weight (in terms of money invested or shares held) to throw around, and you may lack the wealth to buy illiquid, small companies, if that is where you think market mistakes are most often found, since you will not be able to spread your bets. The good news is that you continue to build up your capital, your investment choices will widen, and you can modify your investment strategies accordingly. At the other end of the spectrum, and this is perhaps more the case if you are managing other people’s money, you can have so much capital to invest that some investment strategies become infeasible. For instance, if you are planning on investing in illiquid, small cap stocks, having billions of dollars to invest will increase your transactions costs (by increasing price impact when you trade).

2. Time horizon: Many investors, when asked the question about time horizon, claim to have long time horizons, often because they believe that it is the answer that “good” investors give. The truth is that for most investors, time horizon is as much determined by external factors, such as age, health and liquidity needs, as it is by internal motivations. If you have to pay tuition for your children or expect to have substantial hospital bills in the near future, your time horizon just became shorter, and that has to be factored into your choice of investment strategies.

3. Risk exposure: As with time horizon, the willingness to take risk is partly a function of your personal makeup and partly determined by your life standing. If you have accumulated wealth and have a job with a stable (or rising) income that more than covers your expenses, you are better positioned to take risks than if you are on the verge of retirement, and are investing money that you will be needing soon to cover your post-retirement cash needs.

4. Personal qualities: Your personality and characteristics also come into play in your choice of investment philosophy and strategies. If you are, by nature, impatient, it is unlikely that you will be able to sustain a strategy of buying undervalued companies and waiting for a long time for mistakes to correct. Similarly, if you are easily swayed by peer pressure and what the rest of the world is thinking and doing, it is difficult to be invested in contrarian causes, short-term or long-term. Finally, if your strategy requires special skills to be put into motion, you will have to either have or acquire those skills; a strategy built around finding undervalued companies will require that you know how to value companies and one built around analyzing large and complex datasets looking for mispricing needs statistical and data analysis knowhow.

When investor characteristics and investment philosophy needs are mismatched, there are two negative consequences. The first is that, lacking staying power, investors will abandon strategies well before they should, simply because they are uncomfortable with how they are playing out. The second is that a mismatch creates an emotional cost, where investors struggle with their portfolios and fail what I call the sleep test, where their portfolio’s gyrations keep them awake at night.

Step 5: Keep the feedback loop open

If you have found an investment philosophy that maps on to your market beliefs, found viable strategies that reflect that philosophy and matched it to your personal makeup, you have reached steady state, but only for the moment. That is because almost every part of this process is subject to change, some because of outside forces, and some because of personal changes.

1. Economic setting: Over time, economic settings and structures change, and investment philosophies have to adapt or even be abandoned. For instance, I have argued that technology and disruption have created winner-take-all businesses in the twenty first century, and if you buy into that argument, an investment philosophy (and strategies) built around small cap companies will no longer deliver the payoff it did in the twentieth century.

2. Market lessons: Your views on market mistakes come from looking at data and your own experiences in the market, and as a consequence, they should be revisited as markets change. Just in this century, markets have been tested by crises (the financial crisis of 2008, the COVID meltdown in 2020 and the tariff announcements last April, just to name three), and it is becoming increasingly obvious that assets across classes (stocks, real estate etc) and geographies are moving far more in sync with each other than they did in the last century. That reality has to be integrated into your market views and the investment philosophy/strategies that you use.

3. Trading microstructure: It is undeniable that access to information and trading on most assets has become easier over the last few decades. That is good, but it does come with a cost. Investment philosophies built around the assumption that most investors, especially retail and individual, would not be able to access data or trade easily, may need tweaking, adapting or even abandonment.

4. Personal changes: It won’t come as no secret to you, but you will get older, the amount of capital you have to invest will change, your health and family obligations will shift, and you may even become more or less patient or more or less susceptible to peer pressure. Those factors will all feed into your investment philosophy.

The investing world does not lend itself to absolutes. One of the red flags in investors (retail or institutional) is certitude about their investment choices and views, and an unwillingness to even consider alternatives, a sign that they will be unable to change as the world changes around them.

Book, Class, both or neither?

I like writing, not so much for its commercial potential, but because it allows to get my thoughts in order. I wrote the first edition of my investment philosophies book in ___, and it followed a structure that I have stayed true to, in subsequent editions. I start the book, with a description of what an investment philosophy is and how it first into the investment process, moving on a foundational section, where I look at risk measures, how to read accounting statements and do intrinsic valuation, how transactions costs and taxes drain returns, and at how to test investment strategies that claim to beat the market. In chapters 7 through 12, I spend each chapter looking at a broad investment philosophy (and related strategies), examining evidence for and against each one in the data before outlining what you (as an investor) need to bring to the table to succeed with each one. I close the book, by providing the sobering counter evidence to active investing, where I look at how difficult it to win at that game and the promise and peril of alternative investments (gold, cryptos, fine art, real estate).

If you have one of my earlier editions, is it worth upgrading? If you have the first edition, I do believe it is time, but if you do have the second edition and are budget-constrained, you can hold off. You can find the book online at Amazon and Barnes and Noble, with the latter offering a 25% discount, starting today (March 24).

In parallel, I developed a class that had the same content, and while the NYU certificate version of the class will cost you, I have had a free online version on my webpage, which I created in 2012. That class was in need of an update, and as I finished up the third edition of the book, I created a new version of this class, with forty two sessions covering the same material as the book. Again, if you have taken the earlier version of the class, you may find the material repetitive, but I hope that the updated data and the add ons allow for a richer experience. If you have never taken this class, and online learning works for you, it is designed for investors, individual as well as institutional, and requires little in terms of technical knowledge, and I hope that give it a shot.

The Investing End Game

We all share the same end game in investing, which is to generate the highest returns on the capital we invest, though there are wide variations in how much risk we are willing to take and how long we will wait before cashing out. That is the definition of investment success, but given that there are so many forces that are out of our control, you can do everything right and still fail to meet your objectives, leaving you frustrated and questioning yourself. It is for that reason that a better endgame is to seek out investment serenity, where you end up with an investment path that you are comfortable with, and accept the results that emerge, good or bad.

I have spent this entire post talking about investment philosophies, and in case you have not noticed, I have not shown my hand, on my investment philosophy. I have never believed in hiding behind vague and opaque generalities, and my investment philosophy is built around three principles:

1. Intrinsic value matters: I believe that every asset (anything that generates cash flows) has a intrinsic value, and that with imagination and a willingness to make mistakes, you can estimate that intrinsic value for any company, from start-ups to companies on the verge of default. I believe that much of what passes for valuation in practice is pricing, where people using pricing metrics (such as PE ratios or EV to EBITDA multiples) to make pricing judgments, and that a good valuation requires understanding business models, telling stories and converting these stories into valuation inputs and value estimates.

2. Markets are for the most part right, but make mistakes during periods of uncertainty and change: I never cease to marvel at markets, where millions of individuals with disparate views and information reach consensus on a price. In an age where we have turned over our choices on what movies to watch to Rotten Tomatoes, and which restaurant to eat at to Yelp!, it is worth remembering that markets were the original fount for crowd wisdom. That said, it is also true that markets have provided us with illustrations of crowd madness, where the collective wisdom is hopelessly wrong, and I believe that this is often the case when investors face significant uncertainty, as is the case when companies transition from one stage of the life cycle to another, entire industry groups are faced with the threat of disruption and markets are put into upheaval by crises.

3. Do no harm: While I seek out investments to make that will beat the market, I am cognizant of the reality that I am not entitled to rewards, just because I put in the work, and that luck and chance still can wreak havoc on my best-laid plans. In particular, I have learned, through experience, that my biggest mistakes come from overreach and overactivity, and I have built that learning into my investment philosophy by:

   • Spreading my bets: I have written before about the concentration versus diversification argument, and that what you choose to do as an investor will be a reflection of how much confidence you have in your investment choices, or “conviction”, in investing parlance. I must confess that I don’t share the conviction that concentrated investors bring to the game, and not only spread my portfolio over three dozen stocks, but also follow rigid rules on not letting any single investment exceed 15% of my portfolio.

   • Acting rarely: I don’t trade often, and when I do, I follow the old adage of measuring twice (or three times) before cutting (trading). It helps that I don’t track the market or my portfolio holdings all day, almost never watch the financial news and am not easily swayed by investment sales pitches.

   • Staying away from my weaknesses: I steer away from active market timing and sector bets for a simple reason. I am not good at either, and what I might gain from an occasional win will be wiped out by what I lose in the long term.

   • Being aware of my blind spots: I try to be self-aware, though I don’t always succeed. I know that I am thrown off my game plan by taxes (I don’t like playing them, and that sometimes gets in the way of doing what I should be doing) and I sometimes fall in love with company narratives, because I want them to be true.

This is my philosophy, it reflects my strengths and personality, and it works for me. I sleep well at night and I have no regrets, but I am lucky since I have an clientele of one (or perhaps two) to satisfy. My hope, with both my book and class, is that it provides you with the choices and material for you to find an investment philosophy that works for you and that it delivers the returns you hope to earn, and even if it does not, lets you sleep well at night!

YouTube Video

Investment Philosophies Book

1. Webpage for book

2. Bookseller links

   • Amazon

   • Barnes and Noble (25% off through March 26, 2026)

Investment Philosophies Class

1. My webpage for class

2. YouTube Playlist for class

Status Quo and Standard Practice

Leading into 2008, I had long been skeptical about how we approached the estimation of equity risk premiums, essential ingredients in hurdle rates in corporate finance and discount rates in valuation. It was (and still remains) standard practice to look at historical data, almost entirely from the US, on what stocks had earned over treasuries, and use that historical equity risk premium as the best estimate of the equity risk premium for the future, That approach would have yielded an equity risk premiums of between 5.5% to 14.5%, at the start of 2026, depending on the time period used, the way we compute averages and what we use as the riskfree rate.

These historical equity risk premiums are not only backward-looking and very noisy (see the standard errors), but they allow bias to easily creep in, through the choice of equity risk premiums, with bullish (bearish) analysts picking lower (higher) numbers. Disconcertingly, they also move in the wrong direction, falling during crises (as historical returns get updates) and rising during good times.

A Forward-Looking Alternative

To counter the problems that I saw with historical risk premiums, I started estimating forward-looking equity risk premiums, by essentially backing out from stock prices and expected cash flows, the expected return (internal rate of returns) that markets were pricing into stocks.

That approach yields forward-looking equity risk premiums, and while there is estimation error in the expected earnings growth and payout numbers, it yields vastly more precise estimates that are also model-agnostic. Using this approach, the equity risk premium at the start of 2026 was 4.23% (over the US treasury bond rate):

Note that this estimation is model-agnostic, and is simply a measure of what markets are pricing in, given expected cash flows at the moment.

ERP Estimation during Crises

Unlike historical equity risk premiums, these implied premiums are sensitive to market gauges of fear and greed, and change, as those change. In fact, I computed the ERP, by day, during the 2008 market crisis, and you can see the shifts during that 14-week period below:

Note that the crisis started with the equity risk premiums at 4.2% on September 12, 2008m but almost doubled over the next two months, as stocks went into free fall. To me, these implied equity risk premiums made far more intuitive sense, rising as market fears about banks and the economy rose.

I have continued with the practice of estimating equity risk premiums, by day, during market crises (real or perceived). Here, for instance, is my assessment of the UK market in 2016 in the weeks leading up to the Brexit vote, the market reaction to COVID and the global economic shutdown in 2020, and how the tariffs roiled markets last year. In fact, as we wrestle with an war and oil price induced market shock in March 2026, I started my daily estimates for the ERP on March 1 and will report on how that price has changed over the last two weeks, in the next section.

Equity Risk Premiums - Lessons Learned

The process of estimating implied equity risk premiums on a continuing basis is driven less by intellectual curiosity and more by my need for these numbers, when I value companies. That process has taught me three lessons about equity risk premiums, and I have responded by altering my practices.

1. The equity risk premium is a dynamic and shifting number, and a good estimate of the premium should reflect this volatility. Using an equity risk premium that is different from the implied equity risk premium makes every valuation a joint judgment on what you think about the company and what you think about the market. Put simply, sticking with a 4% equity risk premium during a crisis, when the implied risk premium has surged to 6% will lead you to find most companies to be undervalued, almost entirely because you think that the market is undervalued (not the company). In my view, a company valuation should be market-neutral, and the only way you can get there is by using a current implied equity premium.

My response: Rather than compute the implied equity risk premium at the start of every year, and using that premium over the course of the year, I shifted to computing the equity risk premium for the S&P 500 at the start of every month, in September 2008. I report those numbers on my entry page to my website (damodaran.com) and use them to value companies during the course of the month. You can find these monthly equity risk premium estimates by going to this link.

2. The implied equity risk premium is a consolidated metric for market pricing, and every debate or discussion about whether the market is under or over priced can be reframed as a debate about whether the implied equity risk premium is too low (over pricing), just right (fairly priced) or too high (under pricing). Since the implied ERP incorporates the level of interest rates, expected growth and cash payout, it is a more complete assessment of the market than looking at dividend yields and earnings yields (or variants of PE ratios), two widely used proxies for market pricing. In this post, I took an extended look at how these different measures of equity risk premiums measure up, in terms of predicting future equity returns.

My response: I have been open about my discomfort with timing markets, but when I am asked what I think of the overall market (Is it too high? Is it a bubble?), I first measure the current equity risk premium and then assess it against history. I used this technique to assess US equities at the start of this year in a post, with the accompanying graph:

My conclusion, at the start of 2026, was that while stocks were richly priced using almost every conventional metric (high PE ratios, low dividend yields), the implied equity risk premium was in line with what US stocks have generated over the last 65 years. That said, I did note that 2025 was a tumultuous year, with tariffs making the news and the post-war dollar-centric global economic system starting to fray, and argued that the market seems to be too sanguine about catastrophic risk. Almost on cue, two weeks ago, bombs started falling in the Middle East, and US equities and bonds have been struggling to price in the effects of higher oil prices. In keeping with my practice of estimating equity risk daily, during troubled times, I did compute the implied ERP for the S&P 500 every day, during the last two weeks (Feb 27- March 13):

Oil is up to over a hundred dollars a barrel and the S&P 500 is down, but so far, the market is not behaving as if it is in crisis mode. The equity risk premium, which started March at 4.37% has risen, but only to 4.51%, over the two weeks. In fact, it is the ten-year US treasury bond that has had the bigger surge, up from 3.97% at close of trading, on February 27, to 4.28% at close of trading, on March 13, indicating inflation fears are trumping other market concerns right now. All of this could change next week or the week after, and I will continue to track the equity risk premiums, by day, until the market settles in.

3. The equity risk premium is an essential ingredient into almost every part of financial analysis, incorporated into hurdle rates in corporate finance, discount rates in valuation and in expected returns on equity in financial planning. Given this centrality, I was surprised how little attention it has received from both academics and practitioners, when I looked for references. There is very little usable academic research on equity risk premiums specifically, though there is a great deal on asset pricing and risk. As for practitioners, they have, for the most part, relied on historical risk premiums, and often obtain these premiums from services that summarize the historical data. When I took my first finance class, the historical risk premiums came from data from Ibbotson Associates, that contained annual return data on stocks, bonds and bills. That data was acquired by Duff and Phelps, where it became part of a voluminous book on cost of capital, but much of what that book had to say about equity risk premiums reflected slicing and dicing the historical data, hoping to get further insights, and for the most part failing, because of the noisiness in the data. The US historical data is now in the hands of Kroll, but there is little of value that be extracted by doing deeper and deeper mining expeditions on historical return data. In fact, if you are a fan of historical equity risk premiums (I am not, as you can guess), my suggestion would be to use the Credit Suisse Yearbook, which looks at historical equity risk premiums in 20 markets over more than a hundred years, and does not suffer from the selection bias of focusing on just US data.

My response: I am a practitioner and I decided, for my own understanding, to pull together everything I knew about equity risk premiums into a paper that I wrote in early 2009, and shared online that year. Practitioners seemed to find it useful, and I have updated that paper every year since, at the start of the year. It has grown over time, as I have sought to pull together new findings on equity risk premiums and incorporate changes in markets, and my seventeenth annual update is now ready. I have to confess that at this point, much of the change is data-driven, with tables and graphs updated to include the most recent year’s data, but I hope you still find it useful. The paper resides on the social science research network (SSRN), an Elsevier-run platform for working papers in the social sciences. Unlike most of the other papers on that platform, I have no interest is ever publishing this paper, but you are welcome to download not just the paper, but all of the data that goes with the paper.

Equity Risk Premiums - The 2026 Edition

If you do get a chance to download the paper, I should warn you ahead of time that it long (153 pages), unexciting and entirely directed at practitioners. It is modular, though, and it is broadly broken down into the following sections:

1. The Determinants of Equity Risk Premiums: Given that equity risk premiums represent the price of risk in the market, it should come as no surprise that almost everything that happens in the market, political or economic, affect its level. The picture below summarizes the determinants, and you can find more details in the paper:

As you can see, all of these variables can and will change over time, explaining why the ERP should be a volatile number.

2. Historical Equity Risk Premiums (and spin offs): I spend a section of the paper discussing historical equity risk premiums, examining the statistical properties that make it a faulty approach, and why a belief in mean reversion has made it the status quo. While most of the historical equity risk premiums that you see reported in practice come from the US and are based upon the Ibbotson data going back to 1926, I also look at historical data that goes back further (to 1871) as well as historical premiums in the rest of the world. The historical data on returns in the US has also been mined by services to extract premiums that have been earned by subsets of stocks, and since these premiums often get used by practitioners, I look at the efficacy of these premiums. I specifically look at the small cap premium, a widely used add on in valuation, and not that not only has it been noisy over the entire time period (1926-2025), but that it has disappeared since 1981:

The fact that the small cap premium endures in practice is a testimonial to how once bad practices become embedded in valuation, they never leave.

3. Equity Risk Premiums, by country: While I do have a companion paper that explores country risk in detail, that I update in the middle of the year, I describe my process for estimating equity risk premiums, by country, starting with a mature market premium, and then adding on additional premiums, based on country default risk spreads (based on ratings and sovereign CDS spreads).

4. Implied Equity Risk Premiums and Alternatives: In this section, I start with a description of an intrinsic value model for the market, and use that model to illustrate what you would need to assume for the dividends yield or earnings yield to become reasonable proxies for the equity risk premiums; for the latter, for instance, you have to assume either that there is no earnings growth or that if there is growth, it is value neutral. I then use the full version of the model, allowing for higher growth and cash payout that includes buybacks, to derive my implied equity risk premium estimates. I also look at how my implied equity risk premium estimates relate to other risk proxies (default spreads on bonds, VIX etc.) and how they change over time, as the riskfree rate changes.

5. Efficacy of ERP Estimates: The test of whether an equity risk premium estimate is a good one is in the data, since equity risk premiums measure expectations of what investors hope to earn on equities in future periods. In the last section of the paper, I examine the predictive efficacy of alternative measures of equity risk premiums, by looking at their correlation with actual stock market returns in the next year, the next five years and the next ten years:

Since a good ERP estimate should have a large positive correlation with actual returns on stocks in future years, the current implied premium does best for the five-year and ten-year return, and the historical risk premium does worst, with actual returns increasing (decreasing) when it decreases (increases). In bad news for market timers, none of the equity risk premium approaches does well at forecasting next year’s actual return, and even at the longer time periods, there is significant error in predictions.

Paper

1. Equity Risk Premiums (ERP): Determinants, Implications and Estimates - The 2026 Edition

Data

1. Historical returns on US asset classes (1928 -2025)

2. Implied Equity Risk Premiums, at end of year (1960-2025)

3. Implied Equity Risk Premiums, start of each month (Sept 2008 - March 2026)

4. Equity Risk Premiums, by country, at the start of 2026

Spreadsheets

1. Spreadsheet to compute implied ERP - S&P 500 on February 27, 2026

The Citrini AI Assessment - Report and Responses

The Citrini AI assessment came out on February 22, 2026, and it starts with a preface stating that it is presenting a scenario, not a prediction. I do have issues with that opening, but I will come to them later, but the report itself laid out a story for AI that unfolds with a dark end game for the economy, where by June 30, 2028, the AI disruption has unsettled businesses and displaced workers, with unemployment rates rising above 10% and the market down almost 40% in response. There have been other AI doomsayers, but many of those doomsday scenarios are built around the storyline that AI will not live up to its promise, and the pain comes from having over invested trillions of dollars in building its architecture. In contrast, the Citrini AI story is built on the expectation that not only does AI work well at doing tasks currently performed by white collar professionals, across a range of firms, but its adoption happens very quickly. The pain in the Citrini story comes from that disruption creating substantial job losses, and especially so among higher-earning workers, and the resulting loss of income driving these job losers to cut back on consumption. The ripple effects play out across businesses, with default risks and spreads rising, private credit collapsing and the market and economy pricing in the pain.

I do think that there are major flaws in the steps leading to the economic implosion in the Citrini assessment, but credit should be given where it is due. I have always been troubled by how much we have worshiped at the altar of disruption in this century, putting the founders of disruptors on pedestals and preaching disruption’s virtue. In keeping with Joseph Schumpeter’s description of capitalism as built around creative destruction, I do believe that a vibrant and dynamic economy needs a shake-up and challenging of the status quo, but disruption comes with costs to the businesses that are disrupted, and to the people who work in them. There is much to celebrate, as consumers, in terms of choice and price from the growth of online retail, but that does not take away from the devastation that has been wreaked on brick-and-mortar retail and its constituent parts. Ride sharing has brought car service from its nineteenth century ways into the twenty first century, but at the expense of yellow cabs and conventional car service businesses. The reason that many AI advocates took issue with the Citrini report was precisely because it bought into their sales pitch of how AI bots can not only do what lawyers, bankers, software engineers and consultants do, but also do them better, and then asked the question of “what then?.

The Citrini AI scenario must have hit some targets, because in the days since, we have been flooded with scenarios countering Citrini and arriving at different outcomes. While I was not surprised to see Goldman Sachs, Moody’s and JP Morgan jump in with their AI scenarios, with more benign outcomes for the economy, where the job loss and income effects from AI are modest and temporary, I was surprised to see Citadel wade into the argument, with a direct rebuttal to Citrini, which sees a much more positive end game from AI disruption, and is built around three pillars. The first is the current data on jobs and layoffs in the businesses most directly targeted by AI, such as software, where they note that while jobs have been shed, the job losses have been modest, and AI adoption trends don’t see breakouts consistent with the speedy disruption predicted by Citrini. The second is history, where they look at disruptions in the past (PCs, the internet) and note that none of them have been speedy or have created the job losses or economic collapses predicted in the doomsday scenario. The third is grounded in macroeconomics, where they point to the inconsistency of assuming that a large positive productive shock, from AI’s success, will play out out as large negative shock to the economy and market in which it happens.

Completing the AI story

The problem with all of these AI scenarios is that they are rooted in the weakest of responses to uncertainty, which is to either pick a scenario and to describe it in detail, without establishing, at least in qualitative terms, how likely that scenario is, in the first place, or to list out a whole host of scenarios, without making judgments on likelihood on eany of them. It is entirely possible that what Citrini was presenting was a “worst-case” scenario (I read through the report and could not get a sense of if this was so, and the subsequent responses from Citrini have only muddied the waters), a “low likelihood” scenario or the “likely scenario” of how AI will unfold. If it is a likely scenario, and you buy into the pitch, the investment and personal consequences will be dramatic, since it is entirely possible that, if you are a white-collar worker, you may have lost your job by June 2028, and your savings, if invested in stocks, would have taken a beating. If it is a “low likelihood” scenario, and you are exposed, because of your job, age and portfolio composition, you should consider buying protection, but if it is a worst-case scenario, it is almost entirely useless, except for shock value.

Point Estimates and Probabilities

For much of its history, financial analysis has been built around point estimates, where you identify key drivers, estimate the effects on your bottom line (earnings, cash flows) and make your best judgments. Thus, when valuing a company, you estimate the earnings growth on base year earning, how much you will reinvest of those earnings to grow to get to cash flows, and discount those cash flows back at a risk-adjusted rate to get to value. The problem with point estimates, where almost everything is uncertain is that you will be wrong 100% of the time, though you may still make money, if you are wrong in the right direction.

Financial analysts and economics have been slow in adopting and using probabilistic approaches, where point estimates are replaced by distributions, and a single judgment on outcome by a distribution of outcomes. One reason, at least early on, was that economists and financial analysts often did not have rich enough data or powerful enough tools to use decision trees, simulations or scenario analysis in making their macroeconomic and investment judgments, but that is no longer true. Another reason may be that many in this group are uncomfortable with statistical distributions or probability estimates and stay away from using them, because of that discomfort. The third reason, at least for a subset of analysts, is a concern that being open about estimates and the errors in those estimates, which is visible to all in probabilistic approaches, will be viewed as a sign of weakness or lack of conviction on their part. I have a short paper on using probabilistic approaches, where I look not only at when you may want to use which approach (I look at decision trees, simulations and scenario analysis) but also have a short review of statistical distributions, if you are interested.

Since Citrini specifically titled their AI thought piece as a scenario, I will stick with scenario analysis in this post. In its most sloppy form, and one that has been around for decades, scenario analysis has taken the form of best case - base case - worst case scenarios, an almost useless exercise, since there are almost no risky investments that are going to pass muster under the worst case scenario, no matter how good they are, or are going to fail under the best case scenario, no matter how bad they are. A scenario analysis, done right, should look at scenarios that cover all possible outcomes on an investment or decision, and for completion, need probabilities attached to these scenarios, which can then be used by a decision maker to estimate expected values. That will be almost impossible to do if you are trying to work out future pathways to AI, since it is so early in the process and so little is known about outcomes.

There is an alternate path for scenario analysis that is less information-intensive and thus more feasible, and it draws on the 3P test that I use when valuing companies, where my company valuation narrative has to start with the possible test (it can happen) to being plausible (which requires more backing) and then on to the probable (where you can estimate a likelihood). In the context of scenario analysis, this would require that you categorize scenarios into their the three groupings:

The discussion around where AI is going would become much healthier if scenario proponents were required to state where their proposed scenarios fall in this spectrum. Citrini, for instance, could have saved itself from some of the backlash, if the writer of the AI doomsday report had specified that it was a possible, but not quite plausible scenario.

The AI Disruption - Gaming the Outcomes

In the last week, I have seen at least a dozen scenarios touted by individuals and entities, many of whom I respect, and I must confess that I am whipsawed. If, like me, you are drowning in these scenarios, with very different results and outcomes, the only way to retain your sanity and to take ownership of this process is for you to develop a framework where you can not only put each of these scenarios to the 3P test, but also to develop your own assessment of how AI will play out for businesses, investors and the economy.

1. The Disruption - Form and Speed

The first set of questions that you need to address in the AI story relate to how the AI disruption will evolve, both in form and timing, and to then trace out the aftereffects.

1. AI Disruption Magnitude - Worker Displacement versus Productivity-enhancing Tools: If you listen to some of AI’s lead players, AI will have the capacity to replace workers across multiple businesses, as it develops strengths that go beyond the purely mechanical. One reason that the AI effect on unemployment is so large in the Citrini doomsday scenario is because AI’s reach in the scenario is not just restricted to replacing programmers in software but extends to replacing white collar workers in other technology businesses, financial intermediaries, banking and consulting. In contrast, Citadel’s more benign AI reading comes from AI displacing workers in a smaller subset of businesses, while providing tools in others. At the other end of the spectrum, there are still some who believe that when all is said and done, AI will provide tools to workers that may save them time, but will not be powerful or dependable enough to replace them.

2. AI Disruption Speed: Here again, there is disagreement, with some AI optimists believing that its disruption of regular businesses is imminent, whether displacing workers or in giving them tools. Others believe that AI adoption will take time, partly because the tools need work and partly because businesses and workers are slow to adapt to change. The Federal Reserve in St. Louis has created a tracker of AI adoption rates across users, and while it does not capture the depth of the AI adoption, it does provide a measure of how much familiarity and comfort that users are acquiring, with AI tools.

   

With the caveats about survey data in place, there are interesting trends in these surveys. First, the use of Gen AI tools in non-work settings has grown more than its usage at work, an indication perhaps of how personal devices (phones, in particular) have changed technology adoption rates. Second, the time that AI has saved people, at least so far, has been modest, ranging from less than 1% in the accommodation and food businesses to about 4% in information and management of companies. Overall, this graph suggests that AI usage is neither as explosively fast growing nor as much of a time-saver, as its proponents suggest that it is. The pushback, though, is that these are surveys of the general population, and that there are data points indicating that the disruption effects are more substantial including the substantial write down in market capitalizations of software companies and layoffs at tech companies. The announcement by Block, the fintech company founded by Jack Dorsey, that it would it be letting go of almost 40% of its workforce, for instance, and blaming AI’s rise for the action, was viewed as an indicator of AI’s disruption potential. That is a noisy signal, though, since many tech companies have bloated work forces, and AI gives them easy cover, when correcting past mistakes.

It is true that there is no crystal ball that you can use to gauge the magnitude and speed of AI disruption, but every AI scenario that you see starts with a judgment on one or both.

2. The Disruption Aftershocks

Disruptions create aftershocks, some positive and some negative, and while we often avert our gaze and attention from the latter, a full assessment requires considering both. With AI, the positive effects take the form of higher productivity, as it either allows people to do their jobs more efficiently (with AI tools) or actually replaces people and does their jobs instead, in effect allowing for more output with less labor. Relating back to the different pathways that AI disruption can take, both in form and in form and speed, I would hypothesize that these disruption benefits will be a function of how AI disruption plays out.

Proposition 1: The disruption benefits from AI disruption will be greater from people displacement than from AI productivity tools

Proposition 2: The productivity effects from AI disruption will decrease, at least in economic value terms, the longer it takes for the AI disruption to unfold.

The negative effects of AI, in economic terms, will come from the immediate displacement of people, if AI replaces labor, or from the decrease in employees needed to get tasks done, if AI tools make existing employees more efficient. Here again, I would hypothesize that these disruption costs will be function of how the disruption plays out.

Proposition 3: The disruption costs from AI disruption will be greater from people displacement than from tools, as those laid off lose income and spending power.

Proposition 4: The productivity costs from AI disruption will decrease, at least in economic value terms, the longer it takes for the AI disruption to unfold, since time will allow new entrants into labor markets to adjust to a disrupted business world.

Intuitively, the longer it takes AI to find roots in business, the more time it gives workers time to adjust, retrain or move on. As you can see, the scenarios where AI displaces existing employees and happens quickly are the ones with the biggest benefits and the biggest costs, and the scenarios where AI supplies tools to existing employees and happens slowly has the least benefits and costs. Building on this theme, I see the net effect of AI disruption playing out as follows:

If AI disruption displaces existing workforces, across many businesses, and happens quickly, the net effect is likely to be negative, at least in the near term, since the economy will not only have to absorb major layoffs quickly, but also because those laid off will be higher-earning white collar workers. While that maps on to the Citrini doomsday scenario, there is still much to debate about which industries will see the most job displacement and how quickly these workers will find other jobs. There is also a discussion that should follow, even in this negative net-benefit scenario, of how quickly the economy (and workers) will adapt, and if and whether net benefits will turn positive in the long term. If AI job displacement is on a limited scale, and/or takes time to unfold, both the benefits and the costs of the AI disruption become smaller, but the net benefit is more likely to be positive, in the short and long term. Finally, the AI disruption takes the form of tools that make workers more efficient, but not efficient enough to reduce workforces, both the benefits and costs of AI become much smaller. In fact, if these tools take a long time to craft and displace little or no labor you get the AI disruption fizzle, with very small benefits and costs.

3. The 3P Test

Staying true to my earlier assertion that scenarios without probability estimates are not useful, I will try to put the various AI scenarios that I mapped out in the last section on the 3P continuum.

Let me start with the two possible, but not quite plausible scenarios. The first is the a speedy, massive AI disruption, where AI displaces worker across most businesses, and does so quickly, as visualized by Citrini. It can happen, but given the history of disruption, the limits of AI technology and inertia in the process, it is implausible. At the other extreme, it is possible that AI provides tools to workers that improve productivity marginally, with many ending up being more distractions than tools for productivity, effectively emptying its destructive potential, but that too strikes me as implausible, given what we are seeing in terms of AI capabilities. The most plausible scenarios are ones where AI displaces workers in some industries, such as software and some financial intermediaries, and provides tools that help workers to varying degrees in other businesses. As for probable, I think that disruption will reduce workforces in a subset of businesses, that its tools will include some game changers and that it will take longer to unfold, at least when it comes to monetization, than its advocates think.

My justification for why AI disruption will take time is based on a mix of factors. The first is that my (limited) knowledge and experience with AI products is that while they sometimes work magically well and quickly, they do have kinks, coming partly from being unable to separate good data from bad, and partly from their imperfect attempt to be imitate humans. The second is history, where no disruption has ever unfolded without delays and drawbacks; remember that the dot com disruption almost lost its moorings during the market bust in 2001. The third is human nature, where much as employees and managers claim to want to move on to new and better options, they remain attached to old technology and products; typewriters and mimeographs took a while to disappear after PCs stormed the workplace and flip phones persisted well into the smartphone era.

There are two reasons why I do think that AI disruption is still going to be significant, in the long term. The first is that some of those making the argument that AI will not displace jobs in the long term are assuming that AI in it more advanced form will look like ChatGPT on steroids or be primarily mechanical in its applications. Even my limited exposure to AI’s advanced tools suggests that they have far greater capabilities, and their capacity to mimic human intuition and thought processes is unsettling. The second is the blanket assumption that workers in most white collar jobs will not be easily replaced because they bring training, brainpower and experience into those jobs that will be difficult to replicate. Many white collar workers are bright people with specialized knowledge, but the businesses that hire them put them in straight jackets, pushing mechanics over intuition and rule-driven thinking over principle-driven assessments. In short, it is the nature of the jobs that we have created in many white collar settings that makes them vulnerable to disruption, not the intelligence or training of the people holding those jobs.

It is worth noting that in my probable scenario, AI will unfold at different rates in different businesses, and if I were pushed to distinguish between the businesses that will be targeted most (and soonest) from the businesses where it will take more time, and have less impact, I would look at four factors:

There may be some confirmation and hindsight bias in this table, but there is a good reason why software, a relatively young industry, with young companies and employees, has been one of the first targets for AI disruption. it is profitable, its products and services are logical and rule-based and much of it has no regulatory or system protection. Within software, though, i would expect software that requires more user interface to be more resilient to AI disruption than software that operates in the background. This table, though, can help determine which white collar jobs will be most exposed to AI disruption, and which least, and perhaps also explain why blanket statements about job displacement in banking, consulting and law are overwrought. With banking and law, a substantial portion of the work done is to meet legal or regulatory requirements, not fill operating needs. I have written about the inanity and uselessness of fairness opinions in M&A, where bankers opine on whether an acquiring company is paying a “fair” value for a target, but this practice persists because these fairness opinions provide cover against lawsuits that ensue when deals fall apart. My guess is that the Delaware courts are not quite ready for an AI fairness opinion bot to take the stand and defend a deal, even if the quality of its work is better than a human banker. With consulting, where cookbook solutions are more the norm than the exception, it is worth remembering the clients pay consulting fees not for the advice, but so that they have someone else to blame, when things go wrong, and there too, an AI bot will not have the same outsourcing power as an army of bankers with Harvard MBAs from McKinsey.

4. Cui Bono?

Most of the AI scenarios yield net benefits, and even in the most damaging scenarios, where the AI disruption benefits are overwhelmed by its costs, at least in the short term, you could argue for net positive benefits in the long term. That is good news, but it should taken with a grain of salt, since the distribution of these net benefits across businesses and society will be unequal, and it is possible that the net benefits accrue to a few businesses (and individuals), leaving the rest (businesses and individuals) with net costs.

• The interests of the AI companies and the rest the economy/market will diverge on AI disruption, with the former benefiting if the disruption is across many businesses and happens quickly, and the latter benefiting from a slower disruption restricted to a few businesses. This will be the case even if AI tools add to productivity, since the lower costs that companies acquiring these tools will have as a consequence, may not translate into higher profits, especially if their competitors can pay and acquire the same tools.

• The last few major disruptions, starting with the internet, moving on the China and then the smartphone, have all tilted the playing field in many businesses towards larger companies, making businesses more winner-take-all. It is likely that the AI disruption will play out in similar ways, with the winners winning big, and lots of companies losing out.

• At the individual level, it is not just plausible, but also likely, that a strong AI disruption will make wealth and income inequality worse, with founders of AI businesses joining the ranks of the deca-billionaires and centi-billionaires.

There is one final cost that may not be explicit in economic terms, at least immediately, but one that has to enter the discussions, As AI threatens to displace workers in white collar businesses, it is worth remembering that a job is not just an income-generator, but also a source of self esteem and worth. When software engineers, who pride themselves on their coding skills, bankers, who have spent decades becoming excel ninjas, and consultants, who have found inventive ways of packaging cookbook solutions and presenting them as new and inventive, find that AI can do what they have spent a lifetime perfecting almost effortlessly, the psychic damage will be significant. The fact that blue collar workers lost their jobs to the internet and China disruptions faced a similar predicament and were largely ignored also means that there may be more than a hint of schadenfreude in society’s response to white collar job losses.

The AI Personal Threat

If you are looking at these side costs and threat to jobs that will come from the AI disruption, and wondering whether we should opt out, by regulating or restricting its reach, I am afraid that the choice is out of our hands. The genie is out of the bottle, and the only pathway that you have, if you operate in a space where AI is ubiquitous, is prepare for a reality where AI tools can automate and do much of what you do on a daily basis, but where you have to create a niche or moat that still makes you necessary.

Just about two years ago, I wrote about an AI entity called the Damodaran Bot, that was being developed by Vasant Dhar, my colleague at NYU, and noted that having made all that material that I have developed in my lifetime (classes, books, writing, models, videos) publicly available, I was completely exposed to AI disruption. I have watched that bot develop, with quirks and occasional hiccups, to a point where it can replicate much of what I do almost effortlessly. At the time, though, I did write about what I could do to keep the moat at bay, including the following:

• Generalist vs Specialists: I am a dabbler, an expert in nothing and interested in lots of different things, and I do think that gives me an advantage over a bot that is trained to focus on a topic and drill down. The specialist advantages stem from mastering the vast content in a discipline, but those advantages are diluted with AI entities that can also see that content, but the generalist advantage of using multi-disciplinary thinking with be more difficult for AI to replicate.

• Left and Right Brain: I value companies, and early in my valuation life, I decided that financial modeling was not the right path to value businesses, and that good valuations bridge stories and numbers. If the legend of the right and left brains holds, where the left brain controls logic and numbers and the right brain drives your imagination, a bot will have a tougher time replicating what you do, if you use both sides. That said, I have seen the Damodaran Bot get much better at story telling in the two years that I have watched it, and I need to up my game.

• Reasoning muscle: When faced with questions in the days before the internet, you often had no choice but to reason your way to answers. That may have been time consuming, and your answers might even have been wrong, but each time you did this, you strengthened your reasoning muscles. As we move into a period, where the answer to every question is online, on Google Search and ChatGPT, we are losing the need to exercise those reasoning muscles, and exposing ourselves to being outsourced by our bots.

• An idle mind: I am not a voracious reader nor a listener to podcasts, and since I don’t have much real work to occupy me, I also have plenty of vacant time, with nothing to do. I use that time to daydream and ponder about questions that capture my imagination, including why someone would pay billions of dollars for a sports franchise (like the Washington Commanders), how to deal with the risk of lava from a volcano hitting a spa and ruining its valuation and how streaming has broken the entertainment business. None of these posts include deep insights, but my guess is that the Damodaran bot would have trouble keeping up with my wandering mind.

With the admission that is may not be enough, and that my bot may soon be able write my books and posts, teach my classes and analyze/present data better than I can, I think that you should all be acting as if a bot with your name is looking over your shoulder and trying to learn what you do, and think about what you can do to keep that bot at bay.

There is always the possibility that you are arming yourself for a disruption that fizzles, but I will draw on Pascal’s wager to explain why you should prepare for an AI imitator or bot, even if you don’t believe that it is imminent:

Pascal, a French mathematician, used the wager to explain why be believed in God, even if he was doubtful of a heavenly presence, because the expected value from believing in God exceeded the expected cost from not believing. In the context of AI, acting as if an AI presence and competitor is present will make you better at whatever you do, as a teacher, banker, consultant or software engineer, and that will persist, no matter what AI’s impact is ultimately. Good luck!

YouTube Video

Data Links

1. Federal Reserve in St. Louis AI Adoption Tracker

AI Scenarios

1. The Citrini AI Doomsday

2. The Citadel Response to Citrini

3. Dario Amodei on the Upside (and Dangers) of AI

4. Daron Acemoglu on the Economics (and Adoption) of AI

5. It’s all going to be good - the Jeremy Siegel view on AI

6. The State of AI: Deloitte, McKinsey

The Dividend Decision

I start my corporate finance classes with a description of three core decisions that every firm has to make in the course of business, starting with the investment decision, where you try to invest in projects and investments that earn more than your hurdle rate, moving on to the financing decision, where you decide on the mix of debt and equity to use in funding those investments, and ending with the dividend decision, where firms decide how much cash to return to their owners. In the case of privately owned businesses, this cash can be withdrawn by the owners from the business, but in publicly listed companies, it takes the form of dividends or buybacks. In keeping with the notion that these are the cashflows to equity investors, and that those cash flows should represent what is left after (residual) after all other needs have been met, dividends should reflect that status and, at least in principle, be set after investing and financing decisions have been made:

That utopian view of residual cash being returned to shareholders is put to the test by two real-world realities that often govern corporate dividend policy:

1. Inertia: In many companies, dividend policy is set on auto pilot, with dividends this year set equal to dividends in the last year. It is for that reason that the word I would use to describe dividend policy, at least when it comes to conventional dividends, is ‘sticky’, and you can can see that stickiness at play at US companies, if you track the percentage of companies that increase dividends, decrease dividends or leave them unchanged every year.

   

   In every single year, from 1988 to 2025, the percentage of companies that pay the same dividends that they did in the previous year outnumbers companies that change dividends, and when dividends are changed, they are more likely to be increased than decreased.

2. Me-tooism: In most companies, managers look to peer group dividend policy for guidance on how much, if any, to pay in dividends. Thus, if you are a bank or a utility, it is likely that you will pay high dividends, because everyone else in the sector does so, whereas technology companies will pay no or low dividends, because that is industry practice. While there are good reasons why some industry groups pay more dividends than others, including more predictable earnings and lower growth (and investment needs), hewing to the peer group implies that there will be outliers in each group (fast-growing banks or a mature technology companies) that will be trapped into dividend policies that don’t suit them.

When maintaining or increasing dividends become the end game for a business, you unleash dividend monsters, where investing and financing decisions are skewed to meet dividend needs. Thus, a firm may turn away good investments or borrow much more than it should because it feels the need to sustain dividends.

I have long argued that dividends, in their sticky form, are unsuitable as cash returns to shareholders, but for much of the last century, they remained the primary or often only way to return cash to shareholders. While buying back stock has always been an option available to US companies, its use as a systematic way of returning cash picked up in the 1980s, and in the years since, stock buybacks have become the dominant approach to returning cash for US companies:

As you can see, in the last decade, more than 60% of cash returned to shareholders took the form of buybacks. The primary reason, in my view, is that buybacks, unlike dividends, are flexible, with companies often reversing buybacks, if macro circumstances change, as was the case in 2008 and 2020. There are other reasons that have been offered for the explosive growth in buybacks, but none of them are as significant. There are some who have argued it is stock-based compensation for managers that is pushing them away from dividends to stock buybacks, but that rationale makes more sense for stock options, where stock prices mater, than for restricted stock. In fact, even as more companies shift to restricted stock as their stock compensation mechanism, buybacks have continued to climb, and they are just as high at companies that have no or very low stock based compensation as at companies with high stock-based compensation. Investor taxes are alway in the mix, since investors are often taxed at different rates on dividends and capital gains, but changes in tax law in the last two decades have reduced, if not eliminated, the tax disadvantages associated with dividends, cutting against this argument.

I know that there are many investors, especially in the value investing camp, and quite a few economists, who believe that the shift away from dividends to buybacks is unhealthy, albeit for different reasons. I will return to many of the myths that revolve around buybacks later in this post.

A Rational Cash Return Policy

If you were designing a sensible cash return policy, it has to start with an assessment of how much cash there is available for a firm to return. Since that “potential dividend” should be the cash left over after taxes are paid, reinvestment has been made and debt repaid, it can be computed fairly simply from the statement of cash flows, as free cashflow to equity:

Note that free cash flow to equity starts with equity earnings, converts those earnings to cash flows by adding back depreciation and other non-cash charges, and then netting out capital expenditures and changes in working capital, with increases (decreases) in working capital reducing (increasing) cash flows. It is completed by incorporating the cash flows from debt, with debt issuances representing cash inflows to equity investors and debt repayments becoming cash outflows. Can free cash flows to equity be negative? Absolutely, and it can happen either because you are a money-losing company, too deep in the hole to dig yourself out, or even a money-making companies, with large reinvestment needs? Obviously, paying out dividends or buying back stock when your free cash flows to equity is violating the simple rule that if you are in a hole, you need to stop digging.

If your free cash flow to equity is positive, you can choose to return it to shareholders, either in the form of dividends or buybacks, but you are not obligated to do so. In fact, if you have positive free cashflows to equity and you choose to return none or only a portion of that cash flow, the difference accumulates into a cash balance. If you choose to return more than your free cashflow to equity, you will either have to deplete an existing cash balance, or if you run out of cash, go out and raise fresh capital.

A company that systematically holds back on cash that it could have returned will, over time, accumulate a large cash balance, but that, by itself, may not trigger a shareholder response, if shareholders trust the company’s managers with their cash. After all, cash invested in liquid and riskless investments, like treasury bills and commercial paper, is a neutral (zero NPV) investment, and leaves shareholders unaffected. If you don’t trust management to be disciplined, though, you may punish a company for holding too much cash, effectively apply a “lack-of-trust” discount to the cash. The picture below provides a framework for thinking through the cash return decision, and how it will play out in markets.

As you look at the interplay between earnings, investment needs and potential dividends, you can already see why you should expect cash return policies to change over a company’s life cycle:

The cash returns you see in this graph should largely map on to common sense, with start-ups and very young companies, often money-losing and requiring substantial reinvestment to grow, having negative free cash flow to equity (thus requiring equity infusions). Young growth companies are usually self-funding because internal cash flows may rise to cover reinvestment, but these cash flows are not enough to pay dividends. Mature growth companies have enough cash to return, but stick with buybacks, because they value flexibility. Mature stable companies represent the sweet spot for dividend paying, since they have little in reinvestment needs and large predictable earnings and cash flows. As with everything else in the aging process, companies that refuse to act their age, i.e., young companies that choose to pay dividends or buy back stock or mature companies that insist on holding on to cash, damage themselves and their shareholders.

Dividends in 2025

I will start the assessment of how much companies returned to shareholders in 2025 by looking at conventional dividends paid by companies, using two metrics. The first metric is the dividend payout ratio, where I divide dividends paid by net income, but only if net income is positive; if net income is negative, and dividends get paid, the payout ratio is not meaningful:

As you can see the median payout ratio is about 35% (59%) for US (global) companies, but in both samples, most companies do not pay dividends. There is a sizable subset of companies (12% of US and 14% of global companies) that pay out more than 100% of earnings as dividends, with multiple reasons for that oversized number including a bad earnings year, a desire to increase financial leverage and partial liquidation plans all coming into play.

The second metric is the dividend yield, computed by dividing dividends paid by market capitalization, or dividends per share by the market price per share. In the graph below, I look at the distribution of dividend yields across companies in the graph below, in 2025:

Again looking at only dividend paying firms in the US and global samples, the median dividend yield was 1.10% for the former and 2.43% for the latter, with major divergences across sub-regions; note that the percent of dividend paying firms in the United States has dropped below 30% and even globally, less than half of firms pay dividends. The dividend yield ties into the cost of equity discussion that I initiated in my fifth data update, where I described the cost of equity as the rate of return that investors expect to make on their equity investments. In the United States, for instance, that expected return was about 8.50% at the start of 2026, which would indicate that if you are an equity investor, it is price appreciation that you are dependent on, for the bulk of your equity return.

The dividend yield for equities has declined over time, with the drop off being most noticeable in the United States. The graph below looks at the dividend yield on the S&P 500 from 1960 to 2025, and how that number has become a smaller and smaller portion of the overall expected return on stocks (which I compute with the implied equity return approach):

In 1960, about half of your expected return on stocks came from dividends and that statistic has trended downwards for the last few decades, and in 2025, it represented less than 15% of the total return on stocks.

As a final part of this analysis, I looked at dividend yields and payout ratios, broken down by sector, for both US and global companies:

As you can see, the sectors with the highest percentage of firms paying dividends are financials, real estate and utilities, for both US and global companies, and consumer product companies join in that group, for global companies. In terms of payout ratios, the same three sectors dominate, with energy and real estate returning more than 200% of net income as dividends, in 2025, and posting dividend yields in excess of 6%. Technology companies and communication services have the lowest percent of dividend paying companies and the lowest dividend yields and payout ratios.

The drop in dividend yields over time for the market, the decline in dividend paying firms and the concentration of dividend paying firms in some sectors has put old time value investing to the test. Ben Graham’s strategy of principal protection was built around buying large dividend paying firms and holding on for the long term and it has hit a wall. Any investing strategy built around dividends will result in a portfolio composed of mature and declining firms, and even if you accept that reality, those firms are increasingly concentrated in real estate, banking and utilities.

Buybacks - Myths and Realities

As buybacks have soared in the United States, misconceptions and myths about buybacks have also surged, with some myths used to back up the argument that buybacks are unhealthy and should therefore be banned and others presented as the basis for buybacks as good, representing cannot-lose strategies to beat the market. I will start with the myths that are used to argue against buybacks first, before moving on to those that are used to justify it:

1. Myths in favor of the argument that buybacks are bad and should be restricted or stopped

Myth 1.1: Buybacks are a US phenomenon

Reality 1.1: Buybacks are becoming a global phenomenon

When US firms first started buying back stock in the 1980s, it is true that is was almost entirely or primarily a phenomenon restricted to the US, with large parts of the world restricting or banning the use of buybacks to prevent price manipulation by companies. That is no longer the case, and companies around the world have taken to buybacks, as a flexible alternative to dividends, have adopted the practice. In 2025, I looked at dividends and buybacks from companies around the world:

Companies in the United States are still in the lead in the buyback race, buying back $1.153 trillion in stock in 2025, close to 60% of overall cash returned. Canada, the UK, and Japan are not far behind with more than 35% of cash returned taking the form of buybacks, and the EU and environs, often the slowest to adapt to change, saw almost 29% of cash returned in buybacks. For a variety of reasons, including poor corporate governance and regulatory restrictions, Africa & the Middle East, Eastern Europe and much of south and southeast Asia return relatively little in buybacks.

Myth 1.2: Buybacks are wasteful and reduce corporate investment

Reality 1.2: Buybacks redirect corporate investment from mature companies to growth businesses

The argument that buybacks are wasteful often come from using a firm as a self-contained economic unit, and noting that money used on buybacks cannot be reinvested back into the firm. That is absolutely true, but the cash that goes into buybacks goes to investors and mostly goes back into the market, as investments in other companies. While there are clearly exceptions, where companies that should be investing back into their businesses use that cash to buyback stock, the companies that are the biggest buyers of their own stock are mature firms with insufficient investment opportunities and the companies that have the cash redirected into them need that cash to fund their growth. You can see this play out, when you look at stock buybacks broken down, by age decile (based upon corporate age) for US and global companies:

As you can see, younger companies are not only less likely to buy back stock, but also return less cash in dividends and buybacks, at least as a percent of market capitalization than older companies. Using the life cycle perspective, this suggests that cash is rotating out of older, more mature businesses into younger businesses. I would argue that the difference between geographies where buybacks are rare and geographies where buybacks are common is not in how much corporate investment there is, but in where that investment is directed, with the former investing investing back into declining businesses and the latter funding higher growth and newer businesses.

Myth 1.3: Buybacks are funded with debt are are making companies too highly levered

Reality 1.3: Buybacks are primarily funded with free cash flows to equity and even as buybacks have surged, debt ratios have decreased.

I am not a great believer in case studies precisely because anecdotal evidence is spun into backing priors and preconception.s There are, of course, firms that have dug themselves into a hole by buying back immense amounts of stock, and funding those buybacks with debt, but the aggregate debt ratios for US non-financial service firms, with debt to capital ratios measured against both book and market, have declined over the last four decades, even as buybacks have surged.

If your response is that not all companies buy back stock, and that debt ratios has risen at companies that buy back stock, a comparison of debt ratios (debt to EBITDA and debt to capital) for US firms that bought back stock in 2025 versus those that do not dispels that argument:

If firms are borrowing money to fund buybacks, it is clearly not showing up in the statistics, since companies that bought back stock had much lower debt loads than the companies that did not, a simplistic comparison, but one that carries heft.

Myth 1.4: Buybacks are value-destroying because companies tend to buy back their own stock when prices are too high

Reality 1.4: Buybacks, at any price, can neither add nor destroy value. They can just transfer value

Warren Buffett was late to the buyback party, but when he initiated buybacks at Berkshire Hathaway, he introduced a constraint, which is that he would do buybacks only if he believed that the company’s stock price was less than intrinsic value. He, of course, had the credibility to make this assertion, but most companies don’t impose this constraint and there is evidence that they often buy back their shares when stock prices are higher than they are lower. That does seem like value destruction, but a cash return can neither add nor destroy value, but it can transfer wealth. In the case of stock buybacks at too high a price, wealth is transferred from those who remain loyal shareholders in the firm to those who sell their shares. While there is hand wringing about this, you have a choice, as a shareholder, in a buyback, to sell or hold on, and if you believe that the buyback is at too high a price, you should sell your shares back.

2. Myths in favor of the argument that buybacks are good and generate excess returns for investors

Myth 2.1: Buybacks are value-adding because companies that buy back their own stock when prices are lower than fair value are taking positive net present value investments.

Reality 2.1: Buybacks, at any price, can neither add nor destroy value. They can just transfer value.

This is the inverse of the argument that buybacks are value destroying and they are both grounded in a misclassification of buybacks as projects, rather than cash return, competing with investment projects for the company’s dollars. The truth again is that a stock that buys back stock at lower than fair value is transferring wealth from those who sell back to those who remain, and here again, if you are on the wrong side of wealth transfer, it was your choice to sell back that made you the loser.

Myth 2.2: Buybacks are almost always good for stock prices, since there are fewer shares outstanding after buybacks, and that should increase the price per share.

Reality 2.2: A buyback can increase, do nothing or decrease value per share, depending on the price at which it is done and its effects on leverage.

Buybacks reduce share count (the denominator) but the cash that leaves the firm also reduces fir value (the numerator). A fair-value buyback will create offsetting effects, leaving value per share unchanged, though there can be a secondary effect on value, if the buyback, by reducing equity, changes the debt to capital mix and cost of capital for a company:

It is true that empirical evidence backs up the notion that stock prices benefit from buybacks, but that may be from the selection bias of under levered firms with large cash balances being the biggest players in the stock buyback game.

In general, almost all of these myths come out of treating buybacks as something new and different, rather than a variant on dividends. In general, companies that should not be paying dividends, either because they lack the cash or the future is uncertain, should not be buying back stock either.

Dividend Dysfunction

At the start of this post, I noted that dividend policy is dysfunctional at many firms, driven by inertia (we’ve always paid dividends or we’ve never paid dividend before) and the desire to hew to peer group policies. As a result, there are many companies around the world that adopt dividend policies that, at least of the face of it, take explaining including:

1. Money-losing companies that pay dividends: While there are some companies that offer justifications grounded in worries about sending bad signals or hopes of a bounce back in earnings, many get stuck with dividend policies, because of inertia or peer group pressure, that can drive them into ruin.

2. Money-making companies that refuse to pay dividends: Here again, there can be good reasons for holding back including concerns about whether you can sustain earning and expectations that you will need to invest more in the future, but in some cases, it can unwillingness to initiate dividends in an industry where no one else pays dividends.

3. Negative FCFE companies that return cash (dividends or buybacks): In addition to hopes for a bounce back in FCFE, companies may continue to return cash, even with negative FCFE, because they are trying to increase debt ratios or shrink their businesses over time.

4. Positive FCFE companies that return no cash: Companies that have positive FCFE that don’t return cash may hold back that cash because of the desire to reduce debt ratios or because they ahve investment plans.

The graph below lists out the number of companies in each group, broken down by geography:

Across the globe in 2025, almost 18% of money-losing companies paid dividends, as did about 70% of money-making companies. With FCFE as your indicator, about 37% of companies that returned cash (in dividends and buybacks) in 2025, had negative FCFE, as did 66% of companies with positive FCFE.

Conclusion

There are a whole host of misalignments between what companies return to their shareholders, either as dividends or in buybacks, and what they can, as potential dividends. That suggests to me, and perhaps I am wrong, that investment strategies that are built around cash return, whether they be dividends or buybacks, are likely to go off the tracks. Furthermore, any strategy that is built entirely around dividends, as is the case with strategies where you load up on high dividend yield stocks or buy a handful of heavy dividend payers, such as the Dogs of the Dow, misses the essence of equity investing. A stock is not a bond, where dividends replace coupons, and you get some price appreciation on top, and treating it as such will only create disappointment.

YouTube Video

Data links

1. Dividend statistics, by industry (US and Global)

2. Buyback statistics, by industry (US and Global)

3. Dividends and Buybacks - History for US firms

Spreadsheets

1. Buyback stock price calculator

Data Update Posts for 2026

1. Data Update 1 for 2026: The Push and Pull of Data

2. Data Update 2 for 2026: Equities get tested and pass again!

3. Data Update 3 for 2026: The Trust Deficit - Bonds, Currencies, Gold and Bitcoin!

4. Data Update 4 for 2026: The Global Perspective

5. Data Update 5 for 2026: Risk and Hurdle Rates

6. Data Update 6 for 2026: In Search of Profitability

7. Data Update 7 for 2026: Debt and Taxes

8. Data Update 8 for 2026: Dividends and Buybacks

Debt versus Equity: Choices and Tradeoff

The discussion of the tradeoffs that businesses face on whether to borrow money (debt) or use owner’s funds (equity) has to start with a clear distinction between what it is that sets them apart. While that distinction may seem trivial, since accountants do break financing down into debt and equity on accounting balance sheets, accountants are not always consistent in their categorization, and I think that understanding what sets debt apart from equity can help catch these inconsistencies. There are three dimensions where debt and equity deviate:

1. Nature of claim: Debt gives its holders a contractual claim on the cash flows, insofar as the terms of interest and principal payments are laid down contractually at the time of the borrowing. Note that these contractual claims cover both fixed rate debt, where the interest payments are fixed over the lifetime of the debt, and floating rate debt, where the interest payments will change over time, but in ways that are specified by the bond/loan agreements. Equity gives its holders a residual claim, i.e,, a claim on cash flows, if any, that are left over after other claim holders have been paid.

2. Priority of claim: This follows from the first distinction, but debt holders get first claim on the cashflows, when the firm is in operation, and on liquidation proceeds, if the firm ever goes bankrupt. It is this priority of claims that should generally make debt safer than equity in almost every enterprise that employs both.

3. Legal consequences: A company that fails to pay dividends to its equity investors, no matter how deeply set their expectations of receiving these dividends, may see its stock price drop, but it cannot be held legally accountable for the failure. A company that fails to make its contractual obligations on debt can not only be sued, but can be pushed into bankruptcy, effectively ending its business life.

There are three other distinctions, which do not always hold, but are usually true:

1. Tax Treatment: In much of the world, the tax code is tilted in favor of debt, with interest payments being tax deductible and cash flows to equity (dividends or buybacks) coming out of after-tax cash flows, but there are three caveats. The first is that the tax savings from debt kick in only when a company is generating a taxable profit, though laws on tax loss carry-forwards can allow even money-losing firms to get tax benefits, albeit with a delay. The second is that there are parts of the world, such as the Middle East, where the tax code explicitly bars interest tax deductions, though companies find work arounds sometimes to get the benefits. The third is that there are a few countries that try to even the playing field by either giving a tax deduction to companies for some payments to equity investors (interest on capital as a tax deduction in Brazil) or to investors directly by allowing them credits for corporate taxes paid, when they receive dividends.

2. Role in management: In most businesses, equity investors are given supremacy when it comes to managing the company, exercising that power through either direct ownership or corporate governance mechanisms (such as boards of directors). Again, there are exceptions, as is the case where lenders are given seats on boards of directors or veto power over major operating decisions, but these exceptions are usually triggered when companies violate covenants in loan agreements.

3. Maturity: Debt usually has a finite maturity, though as we saw with the Google hundred-year bond issuance just a few weeks ago, that maturity may be well beyond the lifetime of the buyers of the bond. Equity, in contrast, is, at least on paper, an instrument with no finite due date, and may have cash flows that last into perpetuity.

The figure below captures the differences between debt and equity in the context of a financial balance sheet:

With these distinctions in place, and given that businesses have a choice of using either debt or equity to fund their operations, let us look at the trade off, starting with what the fictional (but often used) reasons for using one source of funding over the other:

One of the most common (bad) reasons that I hear business owners and CFOs of even large companies give for borrowing money is that debt is cheaper than equity. On the face of it, that is of course true, but it is an illusion, at least without the tax benefits kicking in. If the debt is fairly priced, i.e., you are being charged an interest rate that reflects your default risk, borrowing money will make your equity more risky and leave your cost of capital unchanged (if you have no default risk) or raise it (if you have default risk). Intuitively, your cost of capital is designed to capture the risk in your operations, and playing games on the financing side cannot change your operational risk. Among risk-takers, a common reason for using debt is that it will increase your return on equity, and while that again is technically true, it will also raise your cost of equity and magnify the impact of both your successes and your failures. Thus, if you want to borrow money to magnify the payoff to you, as an equity investor, from a successful trade or investment, you should do so, but dispense with the illusion that this is a free lunch. Those who avoid debt have their own share of illusions, starting with the argument that the interest payments on borrowed money will lower net income. That is true, but since you have less equity invested, you may still come out as a beneficiary. They also argue that debt will increase default risk, and lower their bond ratings, but of which are likely to happen, but the objective in business is not to maximize bond ratings, but to increase value; a BBB-rated firm that borrows money and gets tax advantages can be worth more than the same firm with a AAA rating and no debt.

So what are the real trade offs? The first and biggest benefit of debt is its tax treatment, with the tax benefits adding to firm value. Note, and this is said with no moral or ethical judgment attached to it, that this increase in value is coming from taxpayers and not from your operations becoming more valuable. A secondary benefit may come from imposing discipline on managers in public companies, with the need to make interest payments operating as a restraint on a headlong rush into poorly performing investments. On the other side of the ledger, the biggest concern you should have when you borrow money is that it increases the risk of bankruptcy, which if it happens, truncates business life, and even it does not, concerns about it happening can alter how customers, suppliers and investors interact with a business. The other cost that you face when you borrow money is that equity investors and lenders have very different interests, with equity seeking upside and lenders worrying about downside, and the costs of that conflict of interests plays out in covenants and restrictions on operating activity. The figure below summarizes these real trade offs:

The tax benefits versus bankruptcy cost trade off on debt is a simple and very powerful explainer of how much companies should borrow, but in the real world, there are companies that sometimes override the tradeoff and choose to borrow far more or far less than you would expect them to, and they are not necessarily being irrational. Here are three reasons why companies may choose a sub-optimal financing mix:

1. Shields against bankruptcy: If the biggest restraint on borrowing more is the fear of default, anything that reduces or eliminates that fear will cause companies to borrow more money. That default protection can come from governments acting as implicit or explicit guarantors of corporate debt, as was the case with Korean companies in the 1990s, or from seeing other companies in trouble being bailed out by the government, because they were too big to fail.

2. Control versus Value: While businesses have the option of using either equity or debt to fund operations, raising fresh equity usually requires giving up ownership of the business to venture capitalists (at a private business) or to other public market investors (for public companies). For founders and family groups that value control over almost everything else, this can result in firms borrowing money, even though the fundamentals do not support the action. This can explain why Middle Eastern firms, many of which get no tax benefit from debt, may choose to borrow money to fund operations, usually with higher costs of capital, as well as the existence of venture debt, an almost absurd notion from a corporate finance standpoint, since you are lending to start-ups and young money-losing companies with unformed business models and

3. Subsidized debt: If a business has access to debt with below-market interest rates, given default risk, it may make sense to borrow money at these subsidized rates. These debt subsidies are often granted to companies that are seen as delivering on a social purpose (green energy in the last decade) or a political/security interests (defense and infrastructure businesses), and you should therefore not be surprised if they all carry too much debt.

On the other side of the ledger, there are three reasons why companies may borrow less than they should:

1. Restrictive covenants: In markets where debt comes primarily from bankers, it is possible that the covenants that come with this debt are so onerous that businesses will choose to leave tax benefits on the table in order to preserve operating flexibility; this may explain why technology companies, even those with large and stable cash flows, often choose not to borrow money or if they have to, go directly to bond markets.

2. Overpriced equity: Financial markets make mistakes, and sometimes those mistakes may work in your favor as a company with your stock price soaring well above what you think is justifiable, given your fundamentals. In that case, you may choose to use equity, even if you have debt capacity, using your own overpriced shares as currency in funding acquisitions.

3. Regulatory constraints: In some countries and/or sectors, there may be regulatory restrictions on borrowing that cap how much debt you can take on, even though you have the capacity to carry more in debt. Those restrictions can take the form of limits on book debt ratios or on how much interest expense is tax deductible, as a function of revenues or EBITDA.

The picture below captures these frictional considerations:

In sum, the choices between debt and equity play out differently at different companies, depending not only on the characteristics of the company (tax rate, default risk etc.) but also on the management team making that decision on whether to borrow money. If you are an optimizer, by nature, you may this discussion too diffuse, since it points you in a direction (more or less debt) and not to a specific debt mix, but that is easily remedied, if you use the cost of capital as your optimizing tool to find the mix of debt and equity that minimizes your cost of capital.

Download optimizer spreadsheet

Rather than try your patience by belaboring that process, I can point you in the direction of how that is done in my corporate finance class sessions, and with this tool.

Debt and Equity in 2025

With this tradeoff on debt and equity in mind, let’s turn to the data, and in particular, I plan to focus on the choices that companies made globally, on the financing question, in 2025. I will start by looking at the two forces that should have the greatest relevance in this decision, the tax benefits of debt and the default risk, and then look at the mixes of financing across sectors, industries and regions.

The Tax Landscape

Any discussion of taxes has to start with reality checks. The first is that governments need tax revenues, to fund their spending, and corporations and businesses are a target, partly because they affect taxpayers (and voters) indirectly, rather than directly (as is the case with income and sales taxes). The second is that businesses do not like to pay taxes, and try to minimize the taxes they pay, mostly through legal means, with accountants, transfer pricing specialists and tax lawyers abetting, though they sometimes step over the line into tax evasion. When measuring the tax burden that businesses face, we have to distinguish between three measures of tax rates:

1. Marginal Tax Rates: The marginal tax rate reflects the tax rate you face on the last dollar of your taxable income, and thus comes from the statutory tax code of the domicile that the business operates in. While there are a few companies that try to report these tax rates, you are more likely to uncover them by going into the tax code. Fortunately, the leading accounting firms keep updated estimates of these marginal tax rates in the public domain, as do some tax watchdogs, and I used The Tax Foundation for this year’s updates across countries, and the numbers are in the picture below:

Download corporate tax rates, by country

While your eye may be drawn to differences in corporate tax rates, across countries, these differences have narrowed, as the countries with the largest economies (and taxable business) are converging around a marginal tax rate of 25%. There are regional differences, with Latin America and Africa home to some of the highest corporate tax rates, and Eastern Europe and Russia home to some of the lowest. Clearly, there are exceptions within each region, with Ireland the leading outlier in Europe, with a marginal tax rate of 12%, and Paraguay in Latin America, with a marginal tax rate of 10%.

2. Effective tax rates: The effective tax rate is an accounting measure, reflecting the taxes paid and taxable income line items in the income statement, which follows accrual accounting principles. The effective and marginal tax rates can deviate for many reason, including corporate income earned in other countries, tax deferral strategies and even differences between tax and reporting books. I estimated effective tax rates for the companies in my database, and report the averages, by sub-region of the world, in the table below:

   

   In the aggregate, the effective tax rates were lower than the marginal tax rates in about 60% of the companies in my sample, and the difference is a rough proxy for the effectiveness of a tax system, with marginal tax rates running close to or behind effective tax rates in more effective tax regimes. By that measure, India has the least effective tax code among the regions, with an effective tax rate of 22.33% and a marginal tax rate of 30%, followed by the United States and Japan, though the caveat would foreign sales in lower tax locales, in each of these cases. The tax rate statistics, broken down by industry, for global companies, is at this link, if you are interested.

3. Cash tax rates: The cash tax rates also come from accounting statements, with the information in the statement of cash flows used to convert accrual taxes paid to cash taxes paid, and are reflective of what companies actually pay to governments during the course of the year. In 2025, the average cash tax rate across companies with taxable income was 25.86% (21.02%) for global (US) firms, about 1% higher than the effective tax rate in both cases.

For the debt question, it is the marginal tax rate that is most relevant, at least for computing tax benefits, since interest expenses save you taxes at the margin; interest expenses get deducted to get to taxable income, and it is the last dollars of taxable income that thus get protected from paying taxes.

The Default/Distress Landscape

In a world where companies never default, and you still get tax benefits from borrowing, companies push towards higher and higher debt ratios. In the real world, default acts as a brake on debt, with higher default risk translating into lower debt ratios. While default risk is company-specific, the exposure for default risk, across all companies, will vary over time, largely as a function of how well the economy is doing. The ratings agencies (Moody’s, S&P and Fitch) track defaults on a year-to-year basis, and in 2025, they all recorded a drop in default rates across the globe, with US companies driving much of the decline. S&P, in its review of 2025 default and distress, reported that a drop in corporate defaults from 145 in 2024 to 117 to 2025, with the US share of defaults declining from 67% to 62%. To provide historical context, I looked at corporate default rates on loans (using data from FRED) on a quarterly basis going back to 1986:

Corporate loan default rates

While the low defaults in 2025 were a positive sign for lenders, especially given the economic turmoil created by tariffs and trade wars, there were some worrying trends as well. In May 2025, Moody’s estimate of the probability of default at US companies spiked to 9.2%, its highest value since the 2008 crisis. On the bond ratings front, you had more ratings downgrades than upgrades during the year, and almost $60 billion in corporate bonds slipped below investment grade during the year. Breaking down all rated companies, by S&P ratings class, and by region, at the end of 2025

:

Source: S&P Cap IQ

The US has the highest percentage of listed companies with bond ratings, but even in the US, only 11.43% of companies carry that rating, and that percentage is far lower in other parts of the world. Among rated companies, the US has the highest percentage of below investment-grade ratings, suggesting that in much of the rest of the world, there is a self-selection that occurs, where only companies that believe that they will get high ratings are willing to go through the ratings process. Finally, at the start of 2026, there are only AAA rated-companies left in the world, at least according to S&P, in Johnson & Johnson and Microsoft. Looking at 2025, through the lens of default, the numbers look comforting, at least on the surface, with the number of defaults decreasing, but there was disquiet below, as bond buyers wrestled with the consequences of a changing economic world order, and worries about another crisis lurking in the wings.

Debt Burden in 2025

With the background data on tax rates and default risk in place, I will turn to measuring the debt in publicly traded firms, in 2025, and differences in debt burdens across companies, sectors and regions. That mission requires clarity on how to measure debt burdens, and the picture below offers the choices:

Broadly speaking, debt burden metrics can capture debt comfort, i.e., the buffer that businesses have built in to meet their debt obligations and debt level, where you look at debt as a percent of overall funding. In the former group, there are two proxies that you can use to gauge the borrowing buffer - the interest coverage ratio, measuring how much companies have as operating income, relative to their interest expenses, and the debt as a multiple of EBITDA, capturing how many years it will take a company to pay off its debt, if current EBITDA is sustained. In the latter, I will look at debt as a percent of capital invested, using both accounting measures of capital invested (book value) and market value measures.

1. Debt comfort

When companies borrow money, the contractual claims from that debt usually take two forms. The first is interest expenses, and ongoing claim that gives you tax benefits but has to be covered out of income generated each year, and the second is repayment of principal, which comes due at maturity. The interest coverage ratio focuses entirely on the former, and interest payments are scaled to how much a company generates in operating income:

Interest coverage ratio = Earnings before interest and taxes/ Interest expenses

This ratio is simple, with high values associated with less default risk and more safety, at least from a lending perspective. It is still powerful, and it remains the financial ratio that best explains differences in bond ratings across non-financial service companies, and I use it to estimate synthetic bond ratings for firms in my corporate financial analysis.

The problem with interest coverage ratios is that they ignore the other contractual obligation that emerges from debt, which is principal payments due, and the ratio that is most often used to measure that exposure scales total debt at a firm to its earnings before interest, taxes and depreciation:

Debt to EBITDA = Total Debt/ EBITDA

With this ratio, lower values are associated with less default risk and more safety, because a firm, at least if it wanted to, could pay off its debt in fewer years with its operating cash flows.

In the table below, I look at interest coverage ratios and debt to EBITDA values, by sector, for US and global companies, using the same approach I employed in my last update and reporting a ratio based on aggregated values as well as the distribution of the ratio across companies:

As you can see, with both the US and global groupings, technology companies have the largest safety buffers when it comes to debt, with very high interest coverage ratios and low debt to EBITDA, whereas real estate and utilities have the least buffers, with low interest coverage ratios and high debt to EBITDA. As always, the contrast between the aggregated and median values indicate that larger companies, not surprisingly, operate with stronger buffers than smaller companies in almost every sector grouping. Finally, the debt comfort numbers are not computed for financial service companies, for the same reasons that we did not compute costs of and returns on capital for these firms - debt to a bank is raw material and not capital.

2. Debt level

If you go back to the financial balance sheet structure that I started this post with, the debt measure that emerges is one that scales it to the equity invested in the firm (debt to equity) and to the capital invested (debt to capital). These measures have resonance in corporate finance in valuation, because they become drivers of the costs of equity and debt and ingredients in the cost of capital.That said, you can measure this ratio using book value debt to capital (or equity), where you stay with the values of debt and equity reported on accounting balance sheets or with market value debt to capital (and equity ratios), where you use market values for debt and equity. At the risk of sounding dogmatic, book value debt ratios should never come into play in financial analysis and it is market value ratios that matter for two reasons. The first relates back to all of the criticisms I had of accounting invested capital in the context of computing account returns - it is dated and skewed by accounting contradictions and actions. The second is that it is unrelated to what you are trying to measure in a cost of capital, which is what it would cost you to acquire the firm today, where it is market price that determines how much you have to pay, not book value. That said, there remain a fairly large subset of analysts and firms who swear allegiance to book value for a variety of reasons, most of which have no basis in reality. I report book and market debt to capital ratios for all publicly traded firms, broken down by sector for global and US companies:

As you can see, companies look significantly more debt-laden with book value numbers than with market value, and in sectors like technology, where accountants fail to bring the biggest assets on to the books, the difference is even starker. The results in this table reinforce the findings in the debt comfort table, with technology companies carrying very little debt (3-5% in market cap terms) and utilities and real estate carrying the highest. I also reported, on the aggregated numbers, the gross and net debt ratios, with the latter netting cash holdings from debt.

AI Investing and Debt

In every data update post that I have written so far this year, AI has become a component of the discussion, reflecting the outsized role it played not just in market pricing during 2025, but also in business decisions made during the year. To see the connection between AI and debt, I will start with AI investing side, where hundreds of billions were spent by companies building AI infrastructure and large language models (LLMs) during 2025, with plans to spend more in the years to come. A sizable portion of this AI capital expenditure have come from big tech companies, with Meta, Alphabet, Amazon, Oracle and Microsoft all making large bets on the future of AI, and the extent of their investment is visible in the graph below, where I look at capital expenditures and cash acquisitions at these firms (with Broadcom added to the mix) from 2015 to 2025:

The shift at these firms from capital-light to capital-intensive models over this period has been staggering, with the collective investment in 2025 alone hitting $400 billion, with guidance suggesting that they are only getting started. It is worth noting that while big tech has garnered the AI cap ex headline, there are a whole host of other companies that are investing in AI architecture, which include real estate, data centers and power, and many of these companies are still not publicly listed. Going back to investment first principles, you can debate whether these companies can expect to generate positive net present value from their AI investments, and I have argued in earlier posts that it is very likely that they are collectively over investing, with over confidence and a fear of being left behind driving their both corporate investments and investor pricing, in keeping what you would expect when there is a big market delusion.

This big market delusion is a feature, not a bug, and we have seen it play out with dot com stocks in the 1990s, online advertising companies about ten years and even with cannabis stocks in the early years of their listing. The belief that the AI market will be huge, and have two or three big winners, is driving an investing frenzy not just at the big tech companies, but also in smaller start-ups and young firms, but the the market is not big enough to accommodate the expectations across all of these firms, and that will inevitably lead to a correction and clean up.

The AI investing boom enters the financing storyline, which is the focus for this post, because it needs immense amounts of capital. For many of the big tech companies, much of that capital has come from their existing businesses which are cash machines, although the AI cap ex will deplete the free cash flows available to return to shareholders. That said, though, the ramping up of capital investment has been so dramatic that even the cash-rich bit tech companies have turned to debt, as you can see in the graph below:

In 2025, the big tech companies collectively borrowed $160 billion, but given their cashflows and market capitalization, that debt does not put them at risk. For many of the smaller and lower-profile companies investing in this space, where internal cashflows are insufficient, there is a need for external capital, with some coming from equity and a significant portion coming from debt. It is in the context of the debt that I have to pick up on another storyline, which is the rise of private credit as an alternative to banks and the corporate bond market.

As you look at the explosive growth of private credit in this graph, it is worth emphasizing that private credit has been available as an option for borrowers for as long as borrowing has been around, but its usage explode in the last two decades. As AI has increasingly taken a starring role in markets, evidence is accumulating that more private debt is being directed to financing the AI investment boom,. With more than $200 billion in private debt going to AI firms in 2025, AI-related debt is rising as a percent of private credit portfolios.

As private credit has grown as an option, core questions remain of what it brings to a market as differentiating features that allow it to supplant more traditional lending alternatives, i.e. banking and the corporate bond market. Here are some of the reasons offered by private credit advocates for why it may be a preferred choice for entities, in general, and for those investing in AI architecture, in particular:

1. Better default risk assessments: One of the arguments that private credit lenders make is that they have the technical know-how to use data, that banks and bond markets have been more averse to using or have been constrained from using, to get better assessments of default risk. Those assessments, assuming that they are right, allows private credit to lend to entities at rates that are lower than they would be charged, with conventional risk assessments. In principle, that is a solid rationale, but I am unclear about what data it is that traditional lenders are not utilizing that private credit can use, but it is possible that technology and access to the internals of borrowing entities may provide an edge. In fact, the only way to gauge whether this argument of better credit assessment holds up is with a credit shock, where defaults spike across the board.

2. Cashflows-based versus Asset-based lending: A second argument is that traditional lenders, and especially banks, are focused too much on the value of the assets that they are lending against and too little on the cash flows. It is true that bank lending in particular is too focused on asset value, but that focus would provide an opening for private credit in AI, only if AI data centers and architecture investments are poised to start delivering large and positive cash flows soon, and banks are holding back on lending them money. I am hard pressed to think of too many AI investments that have these near-term payoffs.

3. Speedier and more Flexible/Customized Responses: IThis may be the biggest selling point for private credit in the AI investment world, where the investing entities are not just spending billions on AI architecture, but are in a hurry to do so. The regulatory and institutional constraints built into bank lending will stretch the process out in time, and issuing bonds, even if it were an option, comes with its own delay components. In addition, the debt for AI investments may need far more customization than what banks and bond markets can offer, or are allowed to offer, giving private credit an advantage. The problem with speed and customization being the biggest sales pitches for private credit is that it can go with taking short cuts on due diligence and adding terms to loans that cut against prudence, and those can be fatal to lending businesses.

Clearly, these reasons for the presence of private debt have merit, but only to a subset of borrowers, mostly smaller and private, and without a long borrowing history, and for a subset of projects. None that these reasons resonate for the larger tech companies, which have options to borrow money quickly and at fair market rates both from banks and the bond market, and Google’s recent hundred year bond issue is an indication of how much slack bond markets are willing to concede to these firms. When a private credit fund lends Meta for an AI investment, as Blue Owl did in this transaction, the skeptic in me sees either a below-market-rate loan or one with terms that no prudent lender would accept in a loan, and neither is a sustainable lending strategy in the long term.

The coming together of the two storylines on AI and private credit comes with a risk that may extend well beyond the players in these spaces. If you agree with my contention that companies are collectively over investing in AI, driven by the big market delusion, there will be a time when that delusion dissipates and markets will have to correct. In an all or mostly-equity driven space, the pain will be borne by shareholders or owners of these companies, but while painful to them, its ripple effects will be limited. When debt enters the picture, as it has in the AI investment space, the effects of a correction will no longer be isolated to equity investors in these companies, and as private credit gets repriced (from the marking of debt down to reflect higher default risk), the pain to the rest of the economy increases. In effect, we will have a banking crisis created primarily by non-banking lenders behaving badly. We saw some of this start to happen in the last year, as the glow came off the AI rose, and S&P noted the stresses that it put on private credit players. Private credit has had a good run, in terms of delivering returns to investors in it, but it has, in my opinion, the relentless selling of it as an alternative investment class has made it much too big. A shakeout is overdue, which will separate the sloppy lenders from the good ones, and perhaps shrink private credit to healthier levels.

YouTube Video

Data links

1. Marginal and Effective tax rates, by country (January 2026)

2. Debt comfort ratios, by industry (US and Global)

3. Debt load ratios, by industry (US and Global)

Spreadsheet

1. Capital structure optimizer

Paper on Big Market Delusion

1. The Big Market Delusion (with Brad Cornell)

Data Update Posts for 2026

1. Data Update 1 for 2026: The Push and Pull of Data

2. Data Update 2 for 2026: Equities get tested and pass again!

3. Data Update 3 for 2026: The Trust Deficit - Bonds, Currencies, Gold and Bitcoin!

4. Data Update 4 for 2026: The Global Perspective

5. Data Update 5 for 2026: Risk and Hurdle Rates

6. Data Update 6 for 2026: In Search of Profitability

7. Data Update 7 for 2026: Debt and Taxes

The Business End Game

In 1970, Milton Friedman argued in a New York Times article that the social responsibility of a business is to deliver (and increase) profits. That view has come under attack in recent decades, but even in the immediate aftermath of the article’s appearance, there was some push back. Some came from people who argued that Friedman was missing details, with a few noting that it is cashflows, not earnings, that businesses should focus on, and others arguing that it is profits over the long term, not just immediate profits, that should be the focus of a business. My guess is that Professor Friedman would have agreed on both fronts, arguing that he was talking about economic, not accounting, profits, and that there was nothing in his mission statement that foreclosed a focus on long term profits.

In the decades since, there has been a more fundamental critique of the Friedman business end game, coming from those who believe that his view is far too cramped and narrow a vision for a business, and that businesses have obligations to society and the planet that need to be incorporated into decision-making. Initially, these critics argued for imposing social and environmental constraints on the profitability objective, and while Friedman may have taken issue with some of these constraints, arguing that that is what laws and regulations should be doing, he would (probably) have gone along with most of them, given real world frictions. Later, though, these critics decided to go for the jugular, arguing that the business objective itself be reframed to include these broader responsibilities, with some arguing for stakeholder wealth maximization, where businesses seek to maximize value to their different stakeholders (employees, lenders, customers). That idea gained traction among some academics, many of whom never grappled with putting this objective into practice in real businesses, and among some CEOs, who realized that being accountable to everyone effectively meant being accountable to no one, but I am not a fan. About two decades ago, stakeholder wealth maximization was supplemented by ESG, an acronym that quickly got buy-in from the establishment. In 2020, when I first looked at ESG, it was at the height of its allure, with investment managers (led by Blackrock), consultants (with McKinsey up front) and academics, all pushing for its adoption. Given the broad buy in, I expected to see clear and conclusive evidence that ESG was not just good for investors and businesses, but also for society, and I was disappointed on every front. The alpha that was attributed to ESG in investing was accidental, coming almost entirely from its overload on tech stocks in its early years, the evidence that ESG helped businesses deliver higher growth and profits was laughably weak, and on almost every societal dimension that ESG was supposed to make the world a better place, it had failed. Even on risk, the one dimension where a rational argument can be mounted for companies following the ESG rulebook, its impact was hazy, with no discernible effects on costs of capital and only anecdotal (and mostly ex-post) evidence for protecting against reputational and catastrophic risks. In the last five years, ESG has fallen out of favor, largely undone by its own internal inconsistencies, but the gravy train that lived off its largesse has moved on, and taken much of what filled the ESG space, repackaged it, and renamed it sustainability. While advocates for sustainability try to create distance between ESG and sustainability, in my (biased) view, much of that discussion is akin to painting lipstick on a pig and then debating what shade of lipstick suits the pig best, rather than attempting to create real change.

It is with intent, therefor, that I named these three forces - stakeholder wealth maximization, ESG and sustainability - the theocratic trifecta in a post that I wrote three years ago, and argued that they failed for the same reasons.

First, by rooting themselves in virtue rather than in business sense, they rendered a disservice to their own cause. After all, once you decide that you are on the side of goodness, any critics of what you do, no matter how well merited their criticism might be, are quickly consigned to the badness heap, and not just ignored, but also reviled for lacking moral fibre. The problem, of course, is that if an action makes business sense (increases profitability and value), you would not need a virtue brigade to push for that action in the first place. Second, by leaving the definitions of their central ideas (stakeholder wealth, ESG and sustainability) amorphous, they made it easier to sell to investors and companies, but at the expense of consistency and focus. In my 2022 post on ESG, where the Russian invasion of Ukraine had forced its defenders to morph in the face of evidence that that world was more dependent on fossil fuels and defense companies than they had been willing to concede in earlier years, I noted the loss of credibility that comes from shifting definitions of goodness. Third, and most critically, in their zeal to push these concepts to a wider audience and get more people to buy in, they sold a lie, i.e., that you can be good (whatever that definition of good may be) without sacrifice. I have no idea whether ESG and sustainability salespeople meant what they said when they argued that investors could earn higher returns, by adding ESG constraints to their portfolios, and that companies could become more profitable, if they incorporated environmental and social considerations into decision making, but my categorization of people in these spaces as either useful idiots or feckless knaves stems from a refusal to face up to the inherent trade offs.

After decades of pushback from critics of the Friedman business end game, I, for one, believe that Milton Friedman was right, and that we would all be better off to follow up and ask the question of what can be done, given that businesses are profit-seekers, to advance social good and curb externalities. I don’t believe that the disclosure route, which seems to have become the fallback for some seeking better business behavior, will accomplish much, and it may do more harm than good. While laws and regulations can provide a partial fix, they are blunt instruments, and in a setting where businesses can move easily across borders, they may not be effective. Ultimately, we (as consumers and voter) get the businesses we deserve, and if after paying lip service to social causes, we buy products and vote for governments hat undercut those causes, no acronym or word salad will repair the breach.

Profitability in Businesses

I meant to have a short lead-in on why profitability matters at businesses, but as you can see from the previous section, I did get side tracked, but the underlying message is that making money is central to business success and survival, and that measuring profitability is therefore a necessary part of assessing business success and value.

Economic versus Accounting Profits

The Friedman view on the business endgame may have been driven by a vision of economic profits, but in the real world, we are dependent on accounting measures of profits, which are, at best, imperfect substitutes for economic profits. The table below looks at an accounting income statement, highlighting the many measures of profits - gross, operating and net - that you will find in it:

Each profit measure has utility, with gross profits reflecting unit economics, the difference between gross and operating profits capturing economies of scale and the difference between operating and net profits being driven by taxes and choices that businesses make on debt and non-operating assets. In 2025, looking at the aggregate values (in millions of US $) for these line items across sectors, here are the numbers, for both global firms and just the US subset:

In the aggregate, global firms generated $6.2 trillion in net income and $7.7 trillion in operating income on revenues of $72.4 trillion, in 2025; during the same year, US firms generated $2.2 trillion in net income and $2.9 trillion in operating income on revenues of $22.7 trillion in revenues. Across sectors, and looking at revenues, industrials carried the most weight for the global sample, but health care generated the most revenues across the US sub-sample.

Profits scaled to Revenues - Profit Margins

The problem with dollar profits is that comparisons across companies, industries or sectors are skewed by scale differences, and one simple scalar for earnings is revenues, yielding variants of profit margins. While you are undoubtedly familiar with these margin variants, their real use in analysis is in providing insight into business models:

I am not a believer in financial ratio analysis, but I do believe that the income statements for companies, especially examined over time, give us insight into their business models and can help frame valuation narratives. In the table below, I look at differences in margins across sectors in 2025, again looking across global firms, and just US firms:

I have estimated margins, by sector, using the aggregated dollar values for profits and revenues from the previous table, and also reported the cross sectional distribution of company-level margins. Comparing the aggregated margin with the median margin across the sector should give you a sense of how top-heavy the sector is in terms of profitability. In technology, which has the highest weighted operating margin (24.7%) of across sectors, the median operating margin is only 3.41% (-0.30%) across global (US) technology firms; the bigger tech companies are money machines in a sector that still contains a lot of younger and smaller money-losing firms. Note that the margins are not computed for financial service firms, since revenues are often unreported (and mostly meaningless) and gross and operating profits don’t have the same measurement value as they do for non-financial service firms.

Industry Margins and the AI Threat

Breaking down sectors into industries provides more granular detail, and there is a link at the bottom of this post that reports the margin statistics, by industry group. At the risk of stating the obvious, there are large disparities on margins across industries, reflecting differences in unit economics, economies of scale and leverage, as can be seen in this table that lists the industry groupings with the highest and lowest aggregated operating margins among US firms:

At one end of the spectrum, you have industry groups like basic chemicals, which has an aggregated (median) gross margin of 9.31%, making the margin hill much steeper to climb, since operating margins and net margins will be lower. At the other end of the spectrum, in addition to tobacco and railroads (surprised, right?), you have system and application software, delivering an aggregated gross margin of 71.72%, operating margin of 33.21% and net margins of 25.49%, capturing the strong unit economics that characterize the business.

While high margins are a desirable feature for a business, these same high margins can make a business vulnerable to disruption, and the AI sell off that we have seen play out in the last few months in software reflects the concerns that investors have of AI putting significant downward pressure on software margins. If your pushback is that the drop off in revenues and margins has not happened yet, and that it is unfair to software firms to mark their market pricing down preemptively, this is exactly what markets are supposed to do, and these software companies benefited earlier in their lives, when market prices were marked up well ahead of the run-up in margins. You live by the sword (expectations of growth and high margins), you die by it (expectations that growth rates will hit a cliff and margins will decline)!

Time Trends in Profits

I have tracked profit margins for companies for a long time (about three decades) in my datasets, and there is clear evidences that they have trended upwards during the period. In the graph below, I look at the net profit margins for the S&P 500 in the aggregate in this century (from 2000-2025):

As you can see, net profit margins have climbed over the last two decades for US companies, with a number of stories competing for why.

• The most cynical explanation is that this increase in margins is all sleight-of-hand, where accountants are pushing through changes, aided and abetted by accounting rule-writers, to make companies look more profitable. As someone who has taken issue with the gaming of earnings that you often see at companies, I am disinclined to take this criticism seriously, since many of the changes in accounting rules (such as the expensing of stock-based compensation and R&D) should push earnings down, and accountants have more power to move income across periods than they do to increase the level of income.

• A second explanation is that the macroeconomic environment makes it easier for companies to deliver profits, and this explanation had resonance when interest rates were at historic lows in the last decade. As rates have risen back to more normal levels and the economy limps along, I am skeptical of the reasoning in this explanation.

• A third explanation, and this one has been eagerly adopted by many on the political left, is that that this reflects the increase in bargaining power for capital, relative especially to labor, implying that the increase in profits are coming primarily at the expense of worker wages. While there are certainly pockets of the economy where this is true, the margins for most manufacturing and service businesses, which have the highest employee count and wage costs, have stagnated or decreased over the last 20 years, indicating that neither capital nor labor has benefited at least in these sectors.

• The fourth, and in my view the most salient rationale for margin increases, is that the composition of the market has changed, as technology companies supplant old-economy companies, bringing superior unit economics and economies of scale to play. Put simple, a market that gets the largest portion of its value from tech companies will deliver much higher margins that one that gets much of its value from manufacturing and service businesses.

Should we concerned that margins may compress in the future? Of course, and we always should, but that compression, if it happens, will depend almost entirely on how the economy performs and the effects of disruption, if it is coming, for tech companies.

Value Creation in Business

If we define the threshold for business success as generating profits, we are setting the bar too low for a simple reason. Starting a business requires capital, and that capital can earn a return elsewhere on investment of equivalent risk. If those words sounds familiar, it is because I used them in my last post on hurdle rates to describe the costs of equity and capital. Thus, value creation requires a business to generate a return on its equity (capital) that exceeds its cost of equity (capital). That is a simple proposition, and a powerful one, but the measurement challenge we face is in determining the returns that companies generate, and for better or worse, we are dependent on accounting measures of these returns. A good way to see what an accounting return is measuring or at least trying to measure is to look at returns on equity and invested capital in a financial balance sheet:

While accounting returns are widely used in practice, as a gauge of investment quality, they can be skewed not just by accounting inconsistencies but efforts by accountants to do the “right thing” (like writing off bad investments. I have laid out my concerns in exhaustive and incredibly boring detail in this paper on accounting returns, which is dated, but still relevant. I summarize the factors that can cause accounting returns on equity and capital to deviate from reality in the picture below:

With those concerns about accounting returns in place, I computed the accounting returns on equity and invested capital for all of the companies in my global sample (48.156 firms) and my US sample (5994 firms), and the following table reports the statistics for both groups, by sector:

Again, I report the accounting returns computed based on aggregated values first, and then the distributional statistics (first quartile, median, third quartile) for the company-level accounting returns. As with profit margins, you can see that even in sectors where the aggregated accounting returns are high (such as technology and communication services), the median value reflects the reality that most companies in these sectors struggle to deliver double-digit returns.

Turning back to our value creation metric, where we compare accounting returns to costs of equity and capital, you have to be consistent, comparing equity returns to equity costs and capital returns to capital costs

The excess return is a numeric, but as with all numbers in business, it is worth looking behind the number at its drivers, i.e., why do some business deliver returns that consistently outstrip their costs of equity and capital, whereas others struggle? The most powerful explainer of excess returns is not qualitative, since the capacity to generate excess returns comes from barriers to entry and competitive advantages. In the language of value investing, it is the width (strength of competitive advantages) and depth (sustainability of competitive advantage) of moats that determine whether a company can earn more than its cost of equity or capital:

If you are interested in this topic, and it is a fascinating one, Michael Mauboussin brings his erudition and knowledge into play in this Morgan Stanley thought piece from October 2024.

Since I have estimates of costs of equity and capital for each of my firms (see my last data update for details), I compute excess returns, by sector, for my global and US samples:

Given what you saw in the last table, with accounting returns, you should not be surprised to learn that only 29% (28%) of global firms earn returns on equity (capital) that exceed their costs of equity (capital). In fact, if you raise the threshold and look at companies that generate 5% or more as excess returns, the numbers drop off to 19% (17%) for equity (capital) excess returns. Most companies have trouble earning their costs of equity and capital, but if you look at the aggregated values, there are multiple sectors in the US (technology, consumer goods and communication services) that earn double digit excess returns, pointing again to larger companies within these sectors being able to set themselves apart from the rest.

If your concern is that the global statistics are being skewed by regional differences, I compute the excess return statistics broken down by region:

As you can see, there is not a single geography where more than 50% of firms earn more than their required returns, with Japan ranking highest in percentages and Canada and Australia the lowest. Here again, the aggregated values tell a different story, with US companies collectively delivering excess returns of 8.44% on equity and 1.81% on capital, suggesting again that large US companies carry the weight of value creation in the market.

Given how much time we spend in finance examining investments and developing decision rules (NPV>0, IRR>Hurdle rate) that are supposed to protect businesses from taking “bad” investments, you may be surprised at the prevalence of value destroying investments. Some of the failure at businesses to deliver returns on capital that exceed the cost of capital may reflect imperfections in our accounting return measures, since it is based upon earnings in the most recent year, and that may bias us against young and growing companies building up to scale. In my book on corporate life cycle, I highlight how accounting returns shift as companies go from youth to decline

:

To see if this is a factor in our global findings on excess returns, I break companies down by age into deciles and compute excess returns across these groupings:

The table broadly reflects what you should expect to see, with a corporate life cycle, as the percent of companies that beat their cost of capital increase as companies age, but the aggregated excess returns peak in middle age (the middle of the life cycle), more pronounced with US than global firms.

A Profitability Wrap Up

Looking at the data, and there is a danger here that I am overreaching, it seems to me that over the last four decades, moats have crumbled, partly as a result of global competition and partly because of disruption (which upends businesses, turning good businesses to bad ones), and the business landscape has tilted more decisively to larger firms, as more and more businesses become winner(s)-take-all. It is in this context that I take a more jaundiced view of what AI will do for company profitability and value. I believe that, as a disruptor, it will cause downward pressure on margins at most firms, and increase the advantages that larger firms have in each business. How do I reconcile this view with the happy talk of AI as a tool that will make companies more productive, and that the resulting lower costs will make them more profitable? Unless the AI tools that you are talking about are exclusive to these companies, in the sense that competitors cannot buy the same or equivalent tools, these AI tools will lower costs across the board, and competition will then kick in on the pricing front, lowering profitability. If that sounds like a reach, I would recommend a revisit of the US retail sector over the last three decades, as online retail, initially viewed as a boon by brick-and-mortar retail firms, ended up destroying most of them and reducing the margins for retail collectively. As consumers, we will benefit, but as investors or employees in the disrupted companies, we will pay a price that outweigh the benefits, for a sizable number of us. I do think that the AI disruption will be more akin to a slow-motion car wreck, in terms of its effect on overall profitability, and that the margin slippage will occur over time, but it will damaging. Time will tell!

YouTube Video

Datasets

1. Profit margins, by industry (US and Global)

2. Accounting returns and excess returns, by industry (US and Global)

Paper on Accounting Returns (Long and Boring)

1. Return on Capital, Return on Invested Capital and Return on Equity: Measurement and Implications

Data Update Posts for 2026

1. Data Update 1 for 2026: The Push and Pull of Data

2. Data Update 2 for 2026: Equities get tested and pass again!

3. Data Update 3 for 2026: The Trust Deficit - Bonds, Currencies, Gold and Bitcoin!

4. Data Update 4 for 2026: The Global Perspective

5. Data Update 5 for 2026: Risk and Hurdle Rates

6. Data Update 6 for 2026: In Search of Profitability

For a concept tas central to investing and corporate finance as risk is, it is astonishing how much divergence there is across even finance experts and academics on what it is, and consequently on how to measure it. I have heard some describe risk as uncertainty, essentially substituting one fuzzy word for another, others as the threat of grevious loss and still and still others as the possibility of negative outcomes. If you have taken a finance class, and I confess to having a part in this, you may define risk as volatility or standard deviation, or even bring Greek alphabets into play. My favorite definition of risk and one that I start my corporate finance class with is that Chinese symbol for crisis or big risk (and I am sure that I have mangled the symbols, since I have been corrected a dozen times in the past):

As someone who can neither read nor speak Chinese, I am reliant on friends who know the language, and I have been told that the first of the two symbols is the one for danger and the second is a symbol for opportunity. In effect, by bunding together danger and opportunity, the risk measure captures how risk both attracts (to get to opportunity) and repels (with the threat of danger). That duality explains why an investment or business strategy generally cannot be built around the objective of just minimizing risk, since that effectively will remove access to opportunities or recklessly chasing after opportunities, ignoring dangers

With that definition of risk in place, I will start the discussion of risk measures by examining the choices that we face in making the measurement:

1. Upside versus Downside: If you start with a generic definition of risk as receiving an outcome that is different from what your expectation, it is worth recognizing that some of these outcomes will be positive (better than expected) and some will be negative (worse than expected), and that it is the latter than investors and businesses dislike. Thus, there are some who argue that risk measures should focus on just downside outcomes, not all unexpected outcome.

2. Price-based versus accounting-based: Risk measures that are based upon data can be built on market prices, for publicly traded firms, or on accounting data, especially earnings. Price-based measures have the advantage of constant updating, giving you more data, but are sometimes contaminated by the noise and volatility that come from trading. Accounting measures yield more stability, but since they are updated infrequently, and accounting smooths changes over time, they can offer stale or distorted values.

3. Total versus Non-diversifiable: The risk in an investment, whether a project or a business, can come from many different sources, but some of the risks are more investment-specific whereas others are market-wide:

   

To the question of why we should care, the presence of many investments in a portfolio implies that risks that are investment-specific will average out, decreasing or even disappearing as portfolios get larger, whereas market risks remain intact. This insight, which earned Harry Markowitz a Nobel prize, gave birth to modern portfolio theory and is at the heart of most risk and return models in finance.

I have my preferences on how best to measure risk, I would like to keep an open mind and start by laying out the choices we face on risk-measures:

As you can see, the risk measure you choose will be a function of whether you (as an investor or business) believe that the marginal investors, i.e., the investors who own the most shares in your business and trade those share, are diversified or not, and what you believe about financial markets and accounting data.

Risk across Companies in 2025

My sample includes 48,156 publicly traded firms and given that these companies trade across different geographies and are in different businesses, it should come as no surprise that there are wide variations in risk across these companies. In this section, I will start with accounting-based measures, with the caveat that accounting standards vary across the world, though IFRS and GAAP have created significant convergence.

Accounting Measures

While there are a variety of accounting metrics that you can use to measure risk, the most logical one to focus on is earnings, but you have many choices. You could use net income or earnings per share, which will reflect not only the riskiness of the business operate in, but also the amount of debt you have chosen to take on, or you can used operating income, more reflective of just market risk. Within each of these metrics, you can measure risk as volatility (in earnings) or in more simplistic terms, on whether you have positive or negative income. For those investors and businesses to whom, it is debt that is the risk trigger, you can look at measures of that debt burden:

Let’s start with volatility in earnings, where we have two estimation choices that we must make, before we get started. The first is history, and I compute the standard deviations in operating and net income using ten years of earnings data, for each firm, a compromise between a number too high (where I lose too many firms in my sample) and too low (where I lack enough data). The second is that earnings standard deviations in earnings will reflect the level of earnings, with higher earnings companies having higher standard deviations. To control for this, I divide the standard deviation of earnings by the average earnings over the ten years, yielding coefficients of variation in earnings. The following table summarizes the distributional values for this metric, across sectors:

It should come as no surprise that utilities have the least volatile operating earnings and have the lowest coefficient of variation on that metric, and that energy and technology haver the most volatile operating income. On a net income basis, financials and utilities have the lowest volatility in earnings, , and energy and communication services have the highest net income volatility.

If you use the frequency of loss-making, as a risk proxy, the table below captures differences on that metric across sectors on this dimension:

Utilities are again the least risky sector, with a lower percentage of money losers than any other sector, and health care and technology firms have a higher percent of money losers than other sectors.

While there are some who use debt loads as proxies for company risk, and we will come back and look at differences across sectors and industries in a later post, it is a narrow measure, since a young, risky, high growth company with no debt would be classified as low-risk, if it is not debt-laden.

Price-based Measures

All of the stocks in our sample are publicly traded, and consequently, you can use market prices to measure risk. That said, liquidity is a wild card, high in some markets and low in others, and that can cause distortions in the comparison.

1. High and Low Prices: One of the simplest measures of price volatility is the range of prices, with wider divergences between high and low prices at more risky companies and smaller ones at safer companies:

HiLo Risk Measure = (High Price – Low Price)/ (High Price + Low Price)

I computed this statistic for each company in my sample, and then the averages across companies in each industry, and it should be lower (higher) for safer (riskier) stocks. Using my global data, this is what this statistic looks like, across sectors:

Utilities again come in as safest, using this risk metric, tied with real estate, and health care has the widest price ranges of the companies in my sample.

2. Standard deviation in price changes: This is a standard statistical construct, and measures volatility in a stock, though it does not distinguish between upside and downside volatility. Based upon the company-specific standard deviations, again averaged out across sectors, here is what the numbers looked like in 2025:

Financials and utilities are the two safest sectors, and technology and health care are the riskiest, if you measure risk with standard deviation.

3. Betas: If you buy into the notion that the investors setting prices are diversified, and thus care only about risk that cannot be diversified away, you will focus only on the portion of the standard deviation in a stock that comes from the market, and betas, notwithstanding the misinterpretations and misreading, are trying to measure that non-diversifiable portion of standard deviation and scale around one. Again, looking across industries, I look at the distribution of betas, by sector:

If you are interested in a less broad categorization, you can check out betas by industry at the end of this post.

As you review the sector rankings using the varied risk measures, you can see why the heated debates about which risk measure to use is often overdone, since they, for the most part, rank the sectors similarly, with the sectors having less earnings volatility and fewer money-losers also having less volatility in stock price, smaller price ranges and lower betas.

Hurdle Rates
Even as we wrestle with choosing between price and accounting-based measures, it is worth remembering that the end game here is not the risk measure itself, and that risk measures are a means to an end, which is estimating hurdle rates. Hurdle rates come into play for both businesses and investors, setting thresholds that they can use to determine whether to invest or not:

There are some investors and businesses who believe that hurdle rates come from their guts, numbers that reflect personal risk aversion and past experiences, but hurdle rates are opportunity costs, reflecting returns that investors (businesses) can earn in the market on investments of equivalent risk.

In the context of a business, which raises money from debt and equity, you can look at hurdle rates through the eyes of the capital providers – a cost of equity, capturing what equity investor believers expect to make on other equity investments of equivalent risk, and a cost of debt, looking at what lenders can earn on lending to others with similar default risk:

That is what all risk and return models try to do, albeit with different degrees of fidelity to the principle. In fact, my use of an implied equity risk premium in the estimation of the cost of equity is designed to advance this cause, since it is model-agnostic and reflects what investors are pricing stocks to earn, on an annual basis. Thus, when you use the beta in the capital asset pricing model to derive the cost of equity, you should be computing the return you can earn elsewhere in the market on other investments with the same beta, making the cost of equity the hurdle rate for equity investments in a project or company. The cost of capital, which incorporate the cost of borrowing into its construct, is also a hurdle rate, albeit to both debt and equity providers:

As to the question of which of these hurdle rates you should use as a business, the answer lies in consistence. If you are looking at equity returns (return on equity or an internal rate of return based on equity cash flows alone), you should be measuring up against just the cost of equity. Alternatively, with returns on invested capital or an internal rate of return based upon cashflows to the business (pre-debt), it is the cost of capital that comes into play.

I compute the costs of equity and capital for all 48,156 firms in my sample, and in doing so, and in the interests of consistency and ease, I make some simplifying assumptions:

Once I have the costs of equity and capital for each firm, I compute industry averages, both for global firms, and by region (US, Japan, Europe, Emerging Markets, with India and China as sub-categories). You can find the links to the data at the end of this post, but there is another perspective that you can bring to the cost of capital discussion, based upon where a company falls in the company life cycle:

Intuitively, you would expect more uncertainty about business prospects with younger firms, than older ones, especially on the estimation front. That said, it is an open question of whether this uncertainty will translate into higher costs of equity and capital, since it depends on who the marginal investors in these firms are, and whether the risk is diversifiable (and not affect cost of equity) or non-diversifiable. To answer these questions, I classify firms into ten deciles, based on their corporate age, and compute costs of capital:

As you can see, there is no discernible pattern on costs of equity, as you go across the age classes. However, as firms age, they do borrow more, partly because their capacity to generate earnings increase, and that does have some impact on the cost of capital, especially with the oldest firms in the market.
In corporate finance and valuation, an undervalued skill is having perspective, a sense of what comprises typical, and what is a high or a low value. It is for that reason that I also compute a histogram of costs of capital of all publicly traded firms at the start of 2026:

This table is one on my most-used, for many reasons. First, when doing my own valuations, especially for young firms or for firms where the cost of capital is in flux, it gives me the input to us. Thus, if I am valuing a small, AI firm that has just gone public and has global operations, in US dollars, I will start the valuation with a cost of capital of 11.66% and move that cost of capital over time towards 8.65%, as its gets larger and more established. Second, I do see (and must review or grade) other people’s valuations more than I do my own, and this table operates as a plausibility check; a valuation of a publicly traded US company that has a dollar cost of capital of 14% goes on my suspect list, since that is well above the 90th percentile for US firms. Third, the table operates as a reminder that any analysts where the bulk of the time is spent estimating and finessing the cost of capital is time ill-spent, since the 80% of all US (global) companies have costs of capital between 5.26% (6.28%) and 9.88% (11.66%).

For those working in different currencies, the inflation differential approach that I described and used in the last post can be used to convert the entire table. Thus, if you use the expected inflation rates of 2.24% and 4.00% for the United States and India, from the IMF forecasts, you can 1.76% to each of the numbers to each dollar cost of capital that you see in the table or as an industry average.

Conclusion

To run a business or invest in one, you need hurdle rates, and that is what costs of equi6y and debt measure. While models and equations may be how you get these numbers, it is always worth going back to first principles, whenever you face questions on what to do. Thus, recognizing that the cost of capital is an opportunity cost, i.e., the rate of return you can earn elsewhere in the market, on investments of equivalent risk, should be a prompt to use betas that reflect the risk in investments, rather than the entities making the investment, and updated costs of borrowing for the cost of debt. As we enter 2026, we are now in our fourth year with US dollar riskfree rates around 4%, and companies and investors seem to have become acclimatized to the resulting costs of capital, and the shock of seeing dollar riskfree rates surge in 2022, pushing up costs of capital across the board seem to have faded.

YouTube

Datasets

1. Earnings variability, by industry (Global in 2025)

2. Money making and losing percentages, by industry (Global in 2025)

3. Pricing risk measures, by industry (Global in 2025)

4. Betas by industry group (US, Global, Japan, Europe, Emerging Markets, India & China)

5. Cost of capital by industry group (US, Global, Japan, Europe, Emerging Markets, India & China)

Data Update Posts for 2026

1. Data Update 1 for 2026: The Push and Pull of Data

2. Data Update 2 for 2026: Equities get tested and pass again!

3. Data Update 3 for 2026: The Trust Deficit - Bonds, Currencies, Gold and Bitcoin!

4. Data Update 4 for 2026: The Global Perspective

5. Data Update 5 for 2026: Risk and Hurdle Rates

Global Equties in 2025

In my second data update, I noted that US equities had a good year in 2025, delivering a return of 17.72% for the year, but the US dollar weakened in 2025, down a bit more than 7% during the year. I started my exploration of global equities by looking at the returns in local currency terms of equity indices in different parts of the world:

In each region, I have highlighted the best performing index (in green) and worst performing one (in red), and you can see the disparities in market performance, even within regions. One of the problems with comparing returns across currencies is that they are distorted by the effects of inflation that also vary widely across currencies. While I will look at inflation differences in more detail later in this post, one way to make the returns comparable is to recompute them in a common currency. To this end, I compute the dollar returns, in aggregate dollar market capitalization terms, in 2025:

As I mentioned in my second data update, India was the worst performing sub-region of the world, up only 3.31% in 2025, and those returns reflect not just a relatively below-average year in local currency terms, with the Sensex up 8.55% for the year, but a weaker currency, with the rupee depreciating against the dollar. It is only one year and while I will need read too much into it, my argument earlier last year that the India story has legs, but that the path to delivering it will be rockier than many of its advocates seem to thing. For much of the rest of the world, the dollar returns are higher than local currency returns, because of currency appreciation against the dollar.

Zeroing in on the aggregate market capitalization across the world at the start of 2026, I first created a pie chart (on the left) breaking market capitalization by region, and as you can see, US equities, in spite of a weaker dollar, accounted for 47% of global market capitalization.

Evaluating just the change in market capitalization during 2025, in the second pie (not he right), you can see the reason for the slippage in the US hare, with the US punching in below its weight (38% of the change) and Europe and China weighing in, with larger shares.

To close this section, I will unwrite an epitaph for international diversification that many US investors, wealth advisors and market experts were starting to etch in stone even a year ago. For much of the twenty first century, an investor invested entirely in US stocks would have outperformed one who followed the textbook advice to diversify globally. While that may look sound conclusive, the truth is that two decades is not a long time period in stock market history and that you can have extended market runs that look permanent, even when they are not. It is true that as multinationals displace domestic companies, the payoff to international diversification has become smaller over time; buying the S&P 500 would have bought your exposure to the global economy, since the companies in the index, while incorporated in the US, get almost 60% of their revenues in the rest of the world. However, the underperformance of the US, relative to the rest of the world, in 2025 should be a reminder that international diversification still belongs in the toolkit for a prudent investor. That lesson cuts across the globe, and suggests that much as politicians and countries may want to delink from each others, investors don’t have that choice.

Country risk in 2025

If you have been a reader of my posts, I do have a bit of an obsession with country risk,, i.e., why the risk of investing and doing business varies across countries, and what causes that risk to change. My defense for that is that I teach corporate finance and valuation, and to do either, I need answers to these country risk questions, and while you may not like the short cuts and approximations I use along the way, I will take you along on my January 2026 journey:

The place to start any discussion of country risk is with an examination of the factors that feed into that risk, and I will use a matrix that you may have seen in my prior posts on country risk:

While I do take a deeper and more detailed look at these factors in a mid-year update that I do every year (links to paper and my July 2025 blog post), the forces that cause differences in country risk span politics and economics, and include:

1. Political Structure: From an investing and business standpoint, the choice between democracy and autocracy is nuanced, with the former creating more continuous uncertainty, as changes in government bring more policy change , and the latter creating more policy stability in the near term, albeit with a greater likelihood for wrenching and potentially catastrophic uncertainties over time.

2. War and Violence: Investing and business become more hazardous, both physically and economically, if you invest in a more violent setting, and war, terrorism and access to weapons can create differences across countries.

3. Corruption: Corruption affects businesses directly, operating as implicit taxes on businesses that are exposed to it, and indirectly, by undercutting trust and the willingness to follow rules. While differences in corruption across countries are often attributed to cultural factors, a significant component of corruption comes from structures that are designed to encourage and reward it.

4. Legal and Property rights: Investors and businesses are dependent on contracts and legal agreements to operate, but protection for property rights. Legal systems that are capricious in how they enforce contractual and ownership rights, or delay judgments to make them effectively useless, create risks for businesses and investors.

There are many reasons to expect differences across countries, on these dimensions, there is a different perspective that can also help. As some of you may know, I look at businesses through the lens of a corporate life cycle, where as businesses age, their characteristics and challenges change as well. That life cycle structure can be used to explain differences across countries, where the age is less tied to how long a country has been in being and more to do with its economy.

Young economies have higher growth potential, but that higher economic growth comes with more risk (more volatile economies) and require more robust governance to deliver on their promise. As economies age, they face a period of lower growth, albeit with more economic stability, and governance matters less, effectively become mature (middle aged) economies. There is a final phase, where a country’s economy hits walls, and growth can stagnate or even become negative, driven partly by a loss of competitive edge and partly by aging populations. In each of these phases, countries often overreach, with young countries aspiring for the stability of middle age, while trying to grow at double-digit rates, and mature companies, seeking to rediscover high growth. Without treading too much on political terrain, it may be worth thinking about the Trump actions in 2025 as driven, at least partially, by nostalgia for a different time, when the United States was the dominant economic power, with a combination of solid economic growth and stability that few economies, almost unmatched in history.

With that philosophical discourse in country risk out of the way, let’s turn to the brass tacks of measuring country risk, starting with one of the most accessible and widely available one, which are ratings that agencies such as S&P, Moody’s and Fitch (among others) attach to sovereigns. The following is the heatmap of sovereign ratings (from Moody’s) at the start of 2026:

While Moody’s rates more than 140 countries, there remain a few (called frontier markets) that have no ratings, but in terms of the color map, I have included those countries with the lowest rated, because they share many of the same risk characteristics. There are three key features of these ratings that are worth emphasizing:

1. The sovereign ratings are focused almost entirely on default risk, and while the chance that a country will default is correlated with the core risks (violence, political structure, legal system and corruption) that I mentioned up front, there are countries on this list where they diverge. I believe that this is especially the case in the Middle East, where there are countries, like Saudi Arabia, that have low or no default risk, but remain exposed to large political risks.

2. The sovereign ratings have their share of biases, for or against regions, but their bigger sin is that they are slow to react. If you look at the list, you will see countries like Argentina and Venezuela that have seen significant changes in governance and politics in the last year, but where the ratings have not changed or barely changed. That will probably change in 2026, but this delayed response will mean that the sovereign ratings for some countries, at least, will not be good reflections of country risk, at the moment.

3. There were a few ratings changes in 2025, mostly at the margin, but the one that got the most attention was the ratings downgrade for the US that I highlighted at the time it happened. While markets, for the most part, took that ratings downgrade in stride, it did create waves in the process that I use to estimate riskfree rates and equity risk premiums, by country, as you will see later in this post.

The reason that so much of how we deal with country risk rests on sovereign ratings is not because ratings agencies have special insights, but because sovereign ratings, unlike other (often more comprehensive) measures of country risk, like country risk scores (from PRS or the Economist, to name two), can be converted into default spreads that conveniently feed into financial analysis. At the start of 2026, here are my estimates of default spreads for each sovereign rating:

As I noted earlier though, using sovereign ratings to get default spreads comes with the limitations that these ratings may not reflect current conditions, when change is rapid, and that is where the sovereign CDS market has created an alternative. For the 80 countries where sovereign CDS exist, you can get a market-determined number for the default spread, and here are the numbers at the start of 2026:

Note that these spreads, while noisy and reflective of market mood, reflect the world we live in, and both Argentina and Venezuela, which used to be uninsurable, have both seen improvement on these market-driven numbers, albeit from impossible to insure to really costly to insure.

As a final step in my country risk exploration, I repeat a process that I have used to estimate equity risk premiums, by country, every six months for close to three decades. That process starts with estimating an equity risk premium for the S&P 500, and then uses the country default spreads (based upon the ratings) to estimate equity risk premiums for countries:

It is undeniable that the ratings downgrade for the US has created some change in this process. Instead of using the S&P 500’s implied equity risk premium as my estimate of the mature market premium, which was my pathway until May 2025, I now remove the default spread (0.23%) for the US from that premium to get to a mature market equity risk premium (4.23%). To get to country risk premiums for individual countries, I scale up the ratings-based default spreads for the relative riskiness of equities, and add these country risk premiums to the mature market premium:

Download equity risk premiums, by country

Note that I bring the frontier countries into the mix, by using country risk scores for these countries to estimate country and equity risk premiums.

The Currency Effect

While it remains true that country risk and currency volatility/devaluation often go together, one of my concerns with mixing up the two up is that you end up double counting or miscounting risk. To understand the divide between country and currency risk, I start with a look at government bond rates in different currencies, with the caveat that there only about forty governments that issue bonds in their local currencies and that some or many of these government bonds are lightly traded, making their rates unreliable.

In many finance classes and textbooks, you are often taught (as I was) to use the government bond rate as the riskfree rate, on the facile assumption that governments should not default on these bonds, since they can print more currency and cover their debt obligations. The problem with that logic is that it is at odds with the reality that governments can, and often do, default on local currency bonds, choosing that option over devaluation. That also means that the government bond rates can include a default risk component, and to get to a riskfree rate, that default risk needs to be removed from the government bond rate. In the picture above, that is what I do, using the ratings-based default spread). After this clean-up, you can see that riskfree rates vary widely across currencies, from very low in some currencies (Swiss Franc, Japanese yen and the Thai Baht), slightly higher for others (US dollar, Euros) and very high on a few (Turkish Lira, Zambian kwacha).

In my third data update, I estimated an intrinsic riskfree rate for the US dollar, by adding inflation and real GDP growth. Extending that lesson to other currencies, the primary reason for differences in these riskfree rates, across currencies, is expected inflation, with higher(lower) interest rates in higher (lower) inflation currencies. While inflation measures are imperfect and expected inflation estimates are often flawed, I use the IMF’s estimates of inflation to build a global inflation heat map:

The logic that I used to argue that it is unlikely that you will see US treasury bond rates drop much below 4%, at least as long as inflation runs hot (2.5-3%), not only applies for other currencies, but yields a roadmap for estimating riskfree rates in those currencies (including those without a government bond in the local currency). To illustrate, I will try to estimate an Egyptian pound riskfree rate at the start of 2026:

Riskfree rate in local currency = Riskfree rate in US dollars + (Expected inflation rate in local currency – Expected inflation in US $)

Thus, the riskfree rate in Egyptian pounds, using the expected inflation rates of 7.78% for Egypt and 2.24% for the United States is 9.49%:

Riskfree rate in US dollars = US T.Bond rate - US default spread = 4.18% -0.23% = 3.95%

Riskfree rate in EGP (1/1/26) = Riskfree rate in US $ + (Expected inflation in Egypt – Expected inflation in US) = 3.95% + (7.78% - 2.24%) = 9.49%

Note that the riskfree rate in US $ is 3.95%, obtained by cleansing the US 10-year treasury rate on January 1, 2026 (4.18%) of US default risk (0.23%). The estimate for a riskfree rate is an approximation is an approximation, since inflation rates compound, and that compounded version is below:

Riskfree rate in EGP = (1+ US $ Riskfree Rate) × (1 + Expected inflation rate in EGP)/ (1+ Expected inflation rate in US $) -1 = 1.0395 × (1.0778/ 1.0224) -1 = .0958 or 9.58%

I have used IMF inflation rates to get riskfree rates in almost all global currencies in this link, but I don’t blame you, if you are skeptical about the expected inflation numbers. From a financial analysis and valuation perspective, I have good news and it is that it does not matter if you are wrong on inflation, if you are consistently so (in both your earnings and cash flows as well as your discount rates):

Put simply, the effects of expected inflation in valuation cancel out, and that is that the basis of what I would term “the currency invariance theorem”, where the value of a project or company should not change, if you change the currency in which you do your analysis. A project that has a positive NPV, when the analysis is done in US $, should continue to have the same positive NPV, if you redo the analysis in EGP, and a company that is overvalued, when the valuation is in US $, will remain overvalued, if you revalue it in EGP. The currency you chose to do an analysis is cannot alter the underlying value but that does not mean that changes in inflation cannot change the values of businesses, since that effect will depend on how well a company can pass inflation through to its customers (with pricing power), and I examined that relationship in 2022, after inflation had a resurgence in the United States after a decade of being low and boring.

Thus, high inflation in Turkish lira has undoubtedly wreaked havoc the value of some Turkish companies, but given that damage, my point is that revaluing these companies in Euros will not undo that damage.

The Bottom Line

As globalization gets a blowback, and in the midst of turmoil from tariffs, we got a reminder of how, much as we may want to go back to simpler times where the rest of the world did not intrude into our lives, we are all connected in good and bad ways. Thus, you may disagree with me on how to measure country risk and to bring into your analysis and investments, but it is undeniable that risk varies across countries and that we must incorporate that risk into our decision making. I hope that this post expose the layers in the process from the drivers of country risk to how these drivers play out as differences in country ratings, default spreads and equity risk premiums, while illustrating th how country risk can change over time, and sometimes in short periods.

YouTube Video

Datasets

1. Equity risk premiums by country at the start of 2026

2. Differential-inflation riskfree rates, by currency, at the start of 2026

The Trust Narrative

We often underestimate how much of the global economy and financial markets are built on trust - in central banks to preserve the buying power in currencies, in governments and businesses to honor their contractual commitments, in legal systems to enforce them and in norms restraining behavior. That trust can be tenuous, and when violated, not only can the consequences can be catastrophic, but regaining lost trust can be a long, arduous process. In fact, one of the divides between developed and emerging markets for much of the last century was on the trust dimension, with the implicit assumption that emerging countries were less trustworthy than developed countries. That distinction has been muddied in the twenty first century, as crises and political developments have undercut trust in institutions across the board.

I would argue that 2025 was a particularly testing year, as developments in the United States, a dominant player in the global economy and markets, shook trust, and that loss of trust reverberated across its trading partners and global investors.

1. The first of the developments was on the tariff front, where decades of progress towards reducing barriers to trade and establishing predictability was upended on Liberation day(on March 31, 2025), where the US imposed what seemed like arbitrary tariffs on countries, but made those tariffs punitively large. In the immediate aftermath, equity markets around the world went into free fall, and I wrote a post in April 2025 about the tariff effect.

2. Just two weeks later, on April 16, 2025, Moody’s, which had been the lone holdout among the ratings agencies in preserving a Aaa rating for the US, lowered its rating, albeit marginally to Aa1, reducing the number of Aaa rated countries in the world to eight. That rating, though not a complete surprise, still had shock value, and created ripple effects for appraisers and analysts, and I made my assessment in a post in May 2025.

3. On October 1, 2025, the US government went into shutdown mode, as congress balked at increasing the debt limit for the country and on the terms for a new budget, and unlike previous shutdowns, which lasted a few days, this one stretched into weeks, before an agreement was reached to reopen the government on November 12, 2025.

4. In the final months of the year, the independence of the Federal Reserve became a subject of discussion as news stories and pronouncements on social media suggested that the administration was seeking to put its imprint on monetary policy, through its nominees.

Depending on your political persuasion, you may have been one side of the debate or the other about each of these developments, but each of them chipped away at trust in the US government and its institutions.

While Donald Trump is the easy answer to why trust is slipping, the truth is that in each case, the slippage has been occurring over much longer. The push towards uninhibited global trade started running out of steam a decade or more ago, as the costs created political backlash. The Moody’s ratings downgrade followed similar actions by S&P, in 2011, and Fitch, in 2023, partly in reaction to government deficit/borrowing and partly to political dysfunction. The Fed’s much-vaunted independence has always been built more on norms rather that legal strictures, and administrations through the decades have managed to nudge central banks to adopt their preferred paths, and especially so in the aftermath of the pandemic.

The Bond Market

The effect of a loss of trust should be visible most clearly and immediately in the bond market, since bond buyers, of US treasuries, are doing so on the expectation that the US government will not default and that the Fed will do its utmost to preserve the dollar’s buying power (and keep inflation low). Since the shocks from the news stories listed in the section above have the potential to alter both default risk and expected inflation, I looked at the movement of US treasuries over the course of 2025:

US Treasury Data

As you can see, there was little movement in 20-year and 30-year treasuries over the course of the year, but rates dropped, and neither the Moody’s downgrade nor the government shutdown had much effect, and the rise in rates around the downgrade (in April) were more in response to tariffs and preceded the downgrade announcement. In fact, in the face of all of the bad news, the ten-year treasury rate dropped by 39 basis points (from 4.58% to 4.19%) during the year, and short term treasuries dropped even more, effectively altering the slope of the yield curve. To capture that effect, I looked at the evolution of the difference between rates across different maturities over the course of the year:

US Treasury Data

During 2025, the spread between the 10-year and 30-year treasury doubled, the spread between the 10-year and 2-year increased by seven basis points, but at the short end of the maturity spectrum, the spread between the two year and three month treasuries decreased. The net effect was a much more upward sloping yield curve at the end of 2025 than at its start, and while I do not attribute the power to to the yield curve as a prognosticator of future economy growth that some do, it is still marginally a positive sign for the US economy.

To gauge how the news stories played out on the perception of US government default, I looked at the sovereign CDS spreads for the US, market-set numbers capturing the cost of buying insurance against US government default, in 2025:

After a blip in April, where the sovereign CDS spreads increased from 0.4% to just over 0.5% in April 2025, spreads have dropped back to levels lower than they were at the start of the year.

To get a sense of how expectation of inflation changed over the course of the year, I turned again to a market-based number from the treasury market, where the difference between the US ten-year treasury bond rate and the ten-year US treasury TIPs rate (a real rate) operates as a measure of expected inflation:

In 2025, these estimates suggest that the expected inflation barely budged, ending the year lower than it was at the start. That would have put the market at odds with experts, who forecasted a surge in inflation especially after the tariffs were announced, but would have put it in sync with actual inflation reported during the rest of the year.

On the final question of why the Fed independence fight has not created more turmoil in markets, I start with a different perspective from most, since I believe that the role of Fed in setting interest rates is vastly overstated. As I note in that post, the Fed’s much publicized forays into changing the Fed Funds rate has some effect on the short term treasuries, but long term treasuries are driven less by the Fed’s actions (or inaction) and more by expected inflation and real growth. I capture that relationship every year by estimating an intrinsic ten-year riskfree rate, obtained by summing together actual inflation for the year and real GDP growth and comparing it to the ten-year treasury bond rate:

Download intrinsic riskfree data

Over the seventy years of data in this graph, it is clear that the big movements in treasury rates are captured in the intrinsic risk free rate, with higher inflation in the 1970s coinciding with the rise in the treasury rate, and the sustained low rates of the last decade largely in sync with the low inflation and anemic growth during the period. As you can see , after a stint (2021-25) where the intrinsic risk free rate was well above the ten-year treasury rate, largely because of higher inflation, the treasury rate of 4.18%, at the start of 2026, is within reach of the intrinsic rate of 5.10%, obtained by adding inflation and real growth in 2025. That said, though, I do think that the reason that treasury rates stayed well below the intrinsic risk free rate during this period is because markets believed that the Fed would use its powers to try to get inflation under control, even at the expense of a slowing economy (or a recession). It is this belief that will be put at risk if the Fed becomes viewed as an extension of the government, increasing the risks of inflation spiraling out of control, creating a cycle where higher inflation causes higher interest rates, and attempts by central banks to lower these rates actually feed into even higher inflation. It is in the best interests of governments and politicians to let central banks be independent and set rates, because it will lead to better economic outcomes and lower interest rates, while giving politicians cover for unpleasant choices that have to be made to deliver these results.

I complete the assessment of the bond market in 2025 by looking at corporate bonds, and especially at the default spreads of corporate bonds in different ratings classes during the course of the year:

There seems to be a divergence in how the year played out in the corporate bond market, with the higher rated bonds all seeing flat or lower spreads, but bonds below investment grade (below BBB) seeing an increase in spreads.

The Currency Market

Just as bond markets are driven by trust that governments will not default, unless it has run out of options, and that central banks will protect a currency’s buying power, currency markets are swayed by the same concerns. Here, a split emerged between the bond and currency markets. While bond markets, for the most part, took the news stories of the year in stride, the dollar was clearly knocked off balance, and it weakened over the course of the year, as can be seen in the graph below:

The trade-weighted dollar, a broad index of the dollar against multiple currencies, was down 7.24% for the year, but the dollar lost more value against developed market currencies than against emerging market currencies; it was down 8.19% against the former and 6.34% against the latter.

Gold and Silver

When investors lose trust in governments and central banks, it should come as not surprise that their money leaves financial asset markets and goes into collectibles, and in a post in October 2025, I looked at how this played out specifically in the gold market. In 2025, Gold had one of its best years ever, rising 65% during the year, and silver, the other widely held precious metal, had an even bigger year, rising 148% during the year:

The surge in precious metal prices in 2025 was unusual, at least on one dimension. Gold and silver prices tend to rise during periods of unexpectedly high inflation (1970s) or during intense crises, but at least in 2025, neither seemed to be at play. As we noted earlier, inflation came in much tamer than expected, and equity and neither equity nor bond markets, after a brief meltdown in April, showed no signs of trauma. In fact, if you scale gold price to the CPI, the basis for the golden rule, where the argument that gold rises at roughly the inflation rate over time, gold price performance in 2025 broke the indicator, as the ratio of gold price to the CPI exploded well above historic norms.

It is worth noting that a loss of trust in the US government and, by extension, in the US dollar, have translated into increases in gold holdings at central banks, but that increase, while contributing to gold’s allure, cannot explain its price rise during the year. If the rise in gold prices was a surprise, the rise in silver prices was even more so, and in 2025, silver prices rose enough to bring the ratio of gold to silver prices to below the long term median value:

It seems like the market is pulling in different directions on the trust question, with stocks and bonds largely underplaying them, the currency markets indicating some worry and gold and silver suggesting much bigger consequence to the loss of trust. That does not surprise me since the market is not a monolith, and while the broad investor base might have adopted the response of “What, me worry?”, there is a significant segment of investors that see catastrophic risks emerging, and piling into precious metals.

Bitcoin

I have written off and on about bitcoin over the last fifteen years, and have generally straddled the middle, with both sides of the divide (bitcoin optimists and bitcoin doomsayers) taking issue with me. I have argued that bitcoin can be viewed either as a a central-bank free currency, designed by the paranoid for the paranoid, or millennial gold (a collectible), and that we would know better as we saw how it performed in response to macro developments. In many ways, 2025 provided us with a test, which should, if nothing else, advance our understanding of the endgame for bitcoin. In a year where the dollar was weakened as a global currency and central banking independence was questions, you would have expected to see bitcoin do well, both because of its status as a currency without a central bank and as a collectible. The actual price path for bitcoin, in US dollars and Euros, is captured below:

After setbacks in the first third of the year, bitcoin’s price surged upwards in the middle of the year, making those who had built their narratives around it to look good. In my post on bitcoin on July, I focused on the suggestion that other companies should follow the Microstrategy path and put their cash balances into bitcoin, and argued that it was not a good idea. The months following have vindicated that view, as both bitcoin and Microstrategy have seen pricing collapses, and bitcoin ended the year down 6.4% in US dollar terms and 17.4% in Euro terms.

It remains too early in bitcoin’s life to pass final judgment, but if the story for bitcoin is that it will draw in investors who have lost trust in governments and central banks, it is clear that gold and silver were the draws, at least in 2025, not bitcoin. As a final assessment of how the different asset classes moved in relation to each other, I looked at weekly returns in 2025 in six markets - bitcoin, gold, silver, large US stocks, small US stocks and the ten-year treasury bond - and computed correlations across the assets:

There are only a few co-movements which are large enough to be statistically significant. The first is that bitcoin is much more highly correlated with US equities than it is with its collectible counterparts, suggesting that it draws in risk seekers, not the risk averse. The second is that notwithstanding the fact that US treasuries did very little over the course of the year, on a week-to-week basis, their movements affected stock prices. At least in 2025, higher interest rates (translating into negative bond returns) were accompanied by higher stock prices, casting doubt on the notion that the stock market is being held afloat by Fed activity or inactivity.

Conclusion

The big news stories of the year, from the ratings downgrade to the government shutdown to the soap opera of who would lead the Fed all fed into a storyline of fraying trust in US institutions. While that trust deficit should have led to rising interest rates and a tough year for bonds, actual bond market performance, like equities in the prior post, suggested that markets were not swayed. That clearly does not mean that no one cared, since a subset of investors were concerned enough about the trust issue to push the dollar down and put gold and silver prices on stratospheric upward paths. Bitcoin remained the outlier, moving more with stocks and bonds, albeit without their upside (at least this year) and less with collectibles.

YouTube Video

Data Links

1. US Treasury Rates by day in 2025

2. Other Assets (gold, silver, bitcoin), by day, in 2025

3. Intrinsic Riskfree Rates and Treasury Rates from 1954 to 2025

4. Weekly Returns on Asset classes in 2025 (for correlation)

Post links

1. Anatomy of a Market Crisis: Tariffs, Markets and the Economy (April 2025)

2. A Sovereign Ratings Downgrade for the US (July 2025)

3. Fed up with Fed Talk? Fact-checking Central Banking Fairy Tales (October 2024)

4. A Golden Year: Gold Price Surge - The Signal in the Noise (October 2025)

5. To Bitcoin or not to Bitcoin (July 2025)

Data Update links

1. Data Update 1 for 2026: The Push and Pull of Data

2. Data Update 2 for 2026: Equities get tested and pass again!

3. Data Update 3 for 2026: The Trust Deficit - Bonds, Currencies, Gold and Bitcoin!

Back from the Brink: US Equities in 2025

At the start of 2025, the consensus view was that stocks were primed to do well, helped by what investors perceived would be a business-friendly administration and a Federal Reserve, ready to cut rates. In keeping with Robert Burn’s phrase that the best-laid plans of mice and men go awry, the year did not measure up to those expectations at least in terms of policy and rate changes, but stocks still managed to find a way through. Let’s start with a look at the S&P 500 and the NASDAQ, day-to-day through the year:

The first few weeks of 2025 saw of continuation of the momentum built up after the 2024 elections and stock prices continued upwards, but February and March saw a drawdown in stock prices as talk of tariffs and trade wars heated up before culminating in a dramatic sell off in early April, after liberation day, when breadth and magnitude of the tariffs blindsided markets. The sell off was brutal and short, and stocks hit their low point for the year on April 11, 2025. Over the next few months, stocks mounted a comeback, before leveling off at the end of September and coasting for the rest of the year. Early in the year, the S&P 500 held its value better than the NASDAQ, generating talk of a long-awaited tech sell off, but as stocks recovered in the subsequent months, the NASDAQ ended up moving ahead the S&P 500.

Across the entire year, the S&P 500 rose from 5881.6 to 6845.5, delivering price appreciation of 16.39% for the year. The dividends on the companies in the index for the year, based upon dividends in the first three quarters of 2025 and estimates for dividends in the last quarter amounted added a yield of 1.34%.

The S&P 500’s return in 2025 of 17.72% was a solid year, but to provide perspective on how it measures up to history, I looked at annual returns from US stocks from 1928 to 2025, and computed distributional statistics:

Download data

While 2025’s annual returns put it in the right in the middle of the distribution, close to the median and ranked 45th of the 98 years of US equity returns from 1928-2026, it represented a third consecutive year when the annual stock return exceeded the median returns, the longest streak since the mid 1990s; US equities between 2023 and 2025, a period where many market timers were suggesting not just caution but staying out the market, returned 85.32% to investors.

Deconstructing US Stock Price Performance

While stocks had a good year overall, the spoils were dividend unequally, as if often the case, across industries and sectors. To take a closer look at where the best and worst performance was in 2025, I started by looking at a breakdown by sector, where I computed the returns based on the change in aggregate market capitalization in 2025:

I have tracked the performance of each sector, by quarter, and across the year a measured the returns. The best performing sector in percentage returns was communication services (which includes Alphabet and Meta), up 30.63% for the year, followed by technology, which continued it sustained run of success by delivering 23.65% as an annual return; on a dollar value basis, it was not close with technology companies posting an increase of $4.17 trillion in market cap during the year. The worst performing sectors were consumer staples and real estate where the returns were about 2% for the year.

The problem with sector categorizations is the they are overly broad and include very diverse industry groupings, and to overcome that problem, I looked at returns by industry, with a breakdown into 95 industry groups. While you can find the full list at the end of this post, I ranked the industry returns in 2025, from best to worst, and extract the ten best and worst performing industry groups:

Download industry returns in 2025

The surge in gold and silver prices in 2025 carried precious metals companies to the top of the list, with a return of 169.2% for the year, and other energy and mining companies also made the best performer list, with a scattering of technology standouts. The worst performing businesses were primarily old economy, with chemicals, consumer product companies and food processing all struggling during the year.

One of the major changes that we have seen in cross sectional differences in the twenty first century has been the fading or even disappearance of two well documented phenomena from the twentieth century, the first being the small cap premium, where small market cap companies delivered much higher risk-adjusted returns that large market cap companies, and the value premium, where low price to book stocks beat high price to book stocks in the return game. I focused in how these categorizations behaved in 2025, and we did see small cap stocks and low price to book stocks return, at least in part, to favor:

If you are small cap or a value investor, though, I would not be celebrating the return on these premia, but I do think that we will start to see a return to balance, where the groupings will trade off winning in some years for losing in others.

As a final assessment, I did look at the seven stocks that have not only carried the market for the last few years, the Mag Seven, but have been the source of much hand wringing about how markets are becoming top-heavy and concentrated. I started by looking at the individual companies, and how they performed in 2025:

While the Mag Seven saw their collective market capitalization increase by 22.36%, Apple and Amazon lagged with single digit increases, and Nvidia (up 37.8%) and Alphabet (up 62.7%) for the year. Increasingly, the Mag Seven are diverging in their price paths, and that should be expected since they operate in very different businesses and have very different management running them. To examine how much the Mag Seven have carried the market, I tracked the market cap of the Mag Seven against the rest of US equity (close to 6000 companies) from 2014 through the four quarters of 2025.

The aggregate market cap of the Mag Seven has increased from 11% of the US equity market (composed of close to 6000 stocks) in 2014 to 30.89% of the market at the end of 2025, with the $3.9 billion in market cap added in 2025 accounting for 39.3% of the overall increase in market capitalization of all US equities during the year. While this Mag Seven party will undoubtedly end at some point, it did not happen in 2025.

US Equities: Too high, too low or just right?

This post, at least so far, has been a post mortem of the year that was, but investing is always about the future, and the question that we all face as investors, is where stocks will go this year. In my unscientific assessment of stock market opinion, from experts and market timers, there seems to a decided tilt towards bearishness at the start of 2026, for a variety of reasons. There are some who note that having had three good years in a run, stocks will take breather. Others point to history and note that stocks generally don’t do well in the second years of presidential terms. The most common metric that bearish investors point to, though, is the PE ratio for stocks at the start of 2026 is pushing towards historic highs, as can be seen in the graph below, where I look at three variants on the PE ratio - a trailing PE, where I divide the index by earnings in the most recent 12 months, a normalized PE, where I divide the index by the average earnings over the last ten years and a Shiller PE, where I average inflation-adjusted earnings over the last ten years:

Download historical PE ratios for US equities

Using every PE ratio measure, it is undeniable that the PE ratio for the S&P 500, at the start of 2026, is much higher than it has been at any extended period in history, perhaps with the exception with the late 1990s. While this may sound like a slam dunk argument for US stocks being over priced, it is worth remembering that this indicator would have suggested staying out of US equities for much of the last decade. The problem with the PE pricing metric is that it is noisy and an unreliable indicator, and before you use it to build a case that equity investors in the US have become irrational, you may want to consider reasons why US stocks have benefited able to fight the gravitational forces of mean reversion.

1. Robust Earnings Growth & Earnings Resilience: In this century, US stocks have increased more than four-fold, with the S&P 500 rising from 1320.28 at the end of 2000 to 6845.5 at the end of 2025, but it is also worth noting that US companies have also had a solid run in earnings, with earnings increasing about 356% during that same time period.

It is also notable that not only did earnings register strong growth over this period, there were only three years in this century when earnings declined - 2001 (dot com bust), 2009 (2008 crisis) and 2020 (Covid). US companies have become more resilient in terms of delivering earnings through recessions and other crises, pointing to perhaps less risk in equities. I will return in a later post to examine why that may be, with some of the answers rooted in changes in US equity market composition and some in management behavior.

2. Healthy cash returns: In conjunction with delivering earnings growth, US companies have also been returning large amounts of cash to their shareholders, albeit more in buybacks than in conventional dividends. In 2025, the companies in the S&P 500 alone returned more than a trillion dollars in cash flows in buybacks, and in the graph below, I look at how the augmented cash yield (composed of dividends and buybacks) has largely sustained the market:

While the dividend payout ratio, computed using only dividends, has been on a downward trend all through this century, adding buyback to dividends and computing a cash yield ratios yields values that are comparable to what dividend yields used to be, before the buyback era.

In sum, you can see why both bulls and bears retreat to their favored arguments, and there is no obvious tie breaker. The level of stock prices (PE ratios) should be a concern, but you cannot dismiss the benefits of growing and resilient earnings, and substantial cash return. To break the tie, in a very self serving away, I will revert to my favored metric for the US equity market, the implied equity risk premium, which in addition to looking at stock price levels, the growth in earnings and the cash return, also brings in the level of rates. The implied equity risk premium, as I compute it, is the based upon the index level and the expected cashflows (from dividends and buybacks, augmented by earnings growth), and very simply, is an internal rate of return for stocks. Netting out the riskfree rate yields an equity risk premium. The table below contains the computation of the implied ERP at the start of 2026:

Download spreadsheet

Given the index level on January 1, 2026, of 6845.5, and the expected cash flows that I computed on that date (using the dividends and buybacks in the trailing 12 months as my starting point, and growing them at the same rate as earnings), I obtain an expected return on stocks of 8.41%. Subtracting out the US T. Bond rate (dollar riskfree rate) of 4.18% (3.95%) on that day yields an equity risk premium of 4.23% (4.46%) for the US. I want to emphasize again that this estimate is entirely a market-driven number and is model-agnostic.

If you are wondering how estimating this numbers lets you make a judgment on whether US stocks are over priced, all you need to reframe the equity risk premium by asking whether the current ERP is, in your view, too high, too low or just right.

• If you believe that the market is pricing in too low an ERP, given the risks that are on the horizon, you are contending the stocks are over priced.

• If your view is that the current ERP is too high, that is equivalent to arguing that stocks today are under priced.

• If you are not a market timer, you are in effect arguing that the current ERP is, in fact, the right ERP for the market.

To illustrate this point, I have estimated the value of the index at equity risk premiums ranging from 2% to 6%:

With a 2% equity risk premium, you get an astounding value of 14834 for the S&P 500, which would make the index undervalued by 53%. At the other end of the spectrum, with a 6% equity risk premium, the index should trade at 4790, translating into an overvaluation of 43%. So, is the ERP of 4.23% (I will revert to this number, since my historical numbers did use the US treasury bond rate as the riskfree rate) at the start of 2026 a high, low or just-right number? Rather than make that judgment for you, I have computed the implied ERP for the S&P 500 going back to 1960:

Download historical implied ERP

There is something in this graph that almost every investor group can take comfort in, If you are market neutral, you will take comfort from the fact that the current ERP is almost exactly equal to the average for the 1960-2025 period. If you are bearish you will point to the fact that the ERP now is lower than it has been in the post-2008 period, backing up your case that an adjustment is overdue. I am leery of the bubble word, especially used in the context of this market, since unlike the end of 1999, when the ERP got as low as 2.05%, the current ERP is more in the middle of the historic range.

The Bottom Line

US equities had a good year in 2025, and there are signs of excess in at some parts of the market, especially related to AI. That said, the capacity of US companies to continue to deliver earnings and return cash flows even in the face of a tsunami of bad news continues to sustain the market. I am, at my core, a non market-timer, but I have held back on putting idle cash back into US equities in the last year, preferring to keep that cash in treasury bills. It is entirely possible that the market will continue to prove the naysayers wrong and post another strong year, but much as it may pain equity investors, the healthiest development for the market would be for it to deliver a return roughly equal to its expected return (8-9%) and clean up on pricing overreach along the way. For the bears, this may also be the year when the bad news stories of last year, including tariffs and political whiplash, will finally start to hit the bottom line, reducing aggregate earnings and cash flows, but waiting on the sidelines for this to happen has not been a good strategy for the last decade.

YouTube Video

Data Links

1. Historical returns on US equities (also bonds, bills, gold and real estate)

2. Industry returns in 2025, by quarter and industry grouping

3. PE ratios for US equities from 1960 to 2025

4. Historical implied equity risk premium for US equities

Spreadsheet Links

1. Implied ERP computation for the S&P 500 on January 1, 2026