Subscribe now

Something that I am thinking about

Do Hyperscalers (GOOG, MSFT, AMZN) need to permanently hold a “state-of-the-art” (SOTA) model to win?

I have been thinking about this and I think the requirement is becoming less and less necessary. As models become commoditized, the vast majority of economic value accrues directly to hyperscalers’ ability to act as the compute engine, distribution layer, and infrastructure for ALL models at scale.

Why Infrastructure & Scale Dominate SOTA Models

1. The “Toll Road” Strategy (Neutral Platform)

When foundational models become commoditized—driven by open-weights models and rapid architectural improvements (e.g., DeepSeek, Meta’s Llama/Muse series, Mistral, Kimi)—the unit price of raw intelligence drops.

• For pure-play labs (e.g., Anthropic, OpenAI), model commoditization compresses margins because their core product is intelligence. Yes, there might always be a premium placed on the cutting edge and a perceived differentiation, but will this be enough to justify their implied valuations, and will enough volume be willing to pay the premium.

• For hyperscalers (AWS, Azure, Google Cloud), cheaper models trigger Jevons Paradox: as inference becomes cheaper, consumption explodes exponentially. Hyperscalers win by charging a margin on the underlying compute, memory, and networking required to run billions of inference calls daily, regardless of whose weights are running. Hyperscalers can win by commoditizing all of the AI model compliments as GOOG/MSFT/AMZN will serve them all (and charge a small fee 😉.)

2. Inference Economics Are an Infrastructure Game

Serving models at scale is not an easy software solution; it is a brutal physical systems problem:

• Batching and Memory Bottlenecks: Modern reasoning models generate extensive output traces that consume massive high-bandwidth memory (HBM) and Key-Value (KV) cache.

• Utilization Smoothing: Running GPUs at 20% utilization bleeds money. Hyperscalers aggregate global enterprise traffic across different time zones to keep silicon saturated near 100%, driving unit costs per token lower than any single enterprise or middle-tier host can achieve. An analogy is a large international airport with a terminal, many gates, and high volume of throughput- passengers + airplanes. It only works at high utilization, and when it works well it has network effects that become magnetic 🧲.

3. Enterprises Buy a Platform (Resulting in Lock-In)

Enterprises rarely buy raw frontier APIs in isolation. They build on hyperscalers because of:

• Data Residency & Security: Enterprise data stays within AWS VPCs or Azure tenants.

• Integrations & Tooling: RAG pipelines, fine-tuning infrastructure, enterprise search, identity management (IAM), and compliance frameworks are already anchored in the cloud.

So, Why do Hyperscalers Still Pursue a Good Enough In-House Model

While hyperscalers don’t need the absolute #1 model, they still heavily invest in proprietary AI models for two specific reasons:

1. Gross Margin Expansion: Hosting a 3rd-party model (e.g., Microsoft hosting OpenAI) requires sharing revenue. VS Hosting proprietary or open-weights models on custom silicon (e.g., Google running Gemini on TPUs, or AWS running models on Trainium/Inferentia) captures almost 100% of the margin stack.

2. Loss-Leader Distribution: Having a top-tier model acts as the initial magnet that draws developers into their broader cloud ecosystem.

My Conclusion (for the moment)

In the early paradigm of GenAI, model capability was the primary bottleneck. As capability normalizes across open and closed providers, the bottleneck shifts to throughput, latency, energy efficiency, and cost per million tokens.

Hyperscalers are essentially energy-and-silicon utilities. As models commoditize, they morph into multi-model orchestration platforms—monetizing the traffic regardless of which model sits at the top of the leaderboard on any given week.

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , ,

• Housing / Banking - , ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast- Ex-NVIDIA Engineer: Why AI Is About to Get 1000x Cheaper

Neil Movva, co-founder of Sail Research, joins Patrick to explain why the next era of AI may be defined not by faster chatbots, but by background agents that work autonomously for hours, days, or even weeks.

Video- How microSD Cards Store Data

microSD Cards are the most compact and dense data storage format you can buy. Latest microSD Cards can store 2TB in a volume smaller than a penny! A shoebox full of them could hold every movie and TV show on your favorite streaming platform, with room to spare. Watch them open one up and journey down to the nanoscopic scale to see how 16 dies, each as thin as a strand of hair, filled with over 200 layers of NAND flash cells, store all your photos and files.🤩

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Experience Machine: How Our Minds Predict and Shape Reality by Andy Clark Nonfiction, psychology.

• Born to Be Wired by John Malone Autobiography.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏

Capital Markets are Crazy

Most of what I do fits squarely in the tent of long term investing. Munger would call it “sit on your ass and do nothing investing.” Why? Because if the decision is correct at the time of purchase, very little needs to be done after the purchase to improve the growth/results. Go read a book.

But, some of my time is spent profiting from event driven leap option positioning. If, for instance, a company that I happen to know a good deal about goes 20% or more below what I think its fair value is, and I have a favorable long term opinion of the firm, I will open a long term leap option position.

How do we find these opportunities? We have to know and understand a lot of companies, and we often have to make an estimate of what their fair value per share is based on their most recent financials. (Tip- I like to think of the PE as the number of years that I have to wait to be repaid based on the current EPS. 10 years, 20 years, 65 years.) With that in hand, we have to wait and watch for a fat lazy pitch right down the center. Take a look at the table below, and before you continue reading, ask yourself- What looks over/under priced based on its growth prospects?

You already know me, anything with a silly high PE is out. No AMD, no COST. But also anything with a pretty low PEG might be interesting- NVDA, DELL. Everything else (WMT, KO, etc) just goes into the Mr. Market🤪 is crazy pile.

Now, I like to invest in companies that have a cornered resource, or network. Munger would call it a castle with high walls, and a moat around it. I think NVDA is a pretty strong candidate. (This is not investment advice.)

To put a sharper point on it, I can’t believe that a company like Nvidia with the growth prospects that it has, is trading at a price that doesn’t have a much higher PE. Something like this would be a candidate for me to use a Long LEAP Call Option under the expectation that the market discovers its mispricing and eventually recognizes the value.

[Want to read more about LEAP options? I wrote about them in my 2/2025 post, Which Publicly Traded Stocks also have LEAPS Options? Link here.]

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

How Do You Compete With Cloudflare?

Used by roughly 24% of all active websites globally and 28% of the top million busiest websites, Cloudflare has grown in popularity and functionality. Think about that 🤔. About 1 in 4 of every websites that you visit today will be served to you🫵 by Cloudflare… and growing.

As the above table illustrates, Cloudflare does have competition, but it is pretty weak. Each of the names listed above has a slightly different competitive proposition, so in some ways, I am comparing green apples, to red apples, oranges and some grapes. But, this is how the market is presently served. I also have to point out that I think Cloudflare has more products, features, and a larger network, and with that also comes some advantages of scale.

Most companies have new products and feature announcements, but none of them match the pace and frequency of Cloudflare, year over year. Every year Cloudflare hosts a Security Week (March), Developer Week (April), Birthday Week (September), and Innovation Week (October). Birthday week announcements by year for example- 2021 > 20+, 2022 > 36+, 2023 > 20-30, 2024 > 18, 2025 > 25+ . It is impressive to watch.

I am a boring finance guy.🫩 No problem, I am not ashamed of that. And most of these companies are boring CDN Web Server Network companies🥱. They have been professionalized and financialized. (Run like a business by a bunch of adults.) Not Cloudflare. It is run like a fanatic sports fan club. This can be seen from comments and posts from its employees, developers, customers, and not to mention its fanatical leaders Matthew Prince and Michelle Zatlyn.

How do you compete with the energy from a fanatic innovating fan club? How do you compete with Cloudflare? Not every battle is worth it. I didn’t get rich by finding a giant dragon and attempting to slay it. That is nonsense. Better to become friends with the dragon, or quit and run away. Put another way, I think Cloudflare is currently running a great business, and its competitors are un-investable.

Disclosure, I am long Cloudflare, .

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , , ,

• Others- $BRK

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

GPU rental rates still climbing🧗.

Video

Excellent discussion and Ivy is by far one of the best in the housing industry. A conversation with four of the industry’s most influential voices to examine the forces reshaping housing in America. Recorded live at Walker & Dunlop’s Summer Conference. Ivy Zelman, Executive Vice President and Co-Founder of Zelman, a Walker & Dunlop Company, Steven DeFrancis, Founder and CEO of Cortland, Ryan Marshall, President and CEO of PulteGroup, Inc., and Dallas Tanner, President and CEO of Invitation Homes.

[PS- Dream Finders Homes (NYSE: DFH) agreed to acquire Beazer Homes USA, Inc. (NYSE: BZH) on August 7, 2026, in an all-cash transaction.]

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Cinema Speculation by Quentin Tarantino Nonfiction, film. I am a Tarantino fan and it is a fun book. 📖

• The Way of Excellence: A Guide to True Greatness and Deep Satisfaction in a Chaotic World by Brad Stulberg Nonfiction

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏

Steelman🩶 Case for Google to Abandon a SOTA Frontier Model in Favor of Using the Compute for GCP

I have been trying to understand why Google is not as motivated to stay at the frontier, and why they have chosen to reorganize their Deepmind leadership. From a infrastructure and capital efficiency perspective, reallocating massive training clusters away from frontier pre-training and toward Google Cloud Platform (GCP) directly monetizes Google’s constrained compute capacity in a market starved for hardware. Developing cutting-edge frontier models requires tens of billions of dollars in training runs with highly uncertain economic payback, whereas renting out custom TPU clusters and cloud infrastructure yields immediate, high-margin revenue. Enterprise demand for cloud AI infrastructure is outpacing supply; by pivoting compute resources to serve enterprise workloads, hosting third-party models, and supporting custom fine-tuning via GCP, Google transforms volatile research expense into predictable, recurring, high-margin cloud services revenue while avoiding capital expenditures that drag down free cash flow.

Financially and strategically, a “good enough” model strategy aligns with actual market consumption, where the marginal value of frontier intelligence is rapidly diminishing relative to cost. The vast majority of enterprise use cases—such as document parsing, code completion, agentic routing, and customer support—do not require a trillion-parameter flagship model; they demand low latency, tight security, and low per-token pricing. As intelligence commoditizes and smaller distilled or specialized models achieve near-frontier benchmarks at a fraction of the inference cost, spending exponential capital for minor benchmark gains creates a negative ROI. Deploying performant, cost-effective models allows Google to capture the bulk of enterprise volume, preserve gross margins on inference, and leave competitor labs to burn capital chasing marginal gains at the bleeding edge.

And lest we forget, Google also owns ~14% of Anthropic. I believe we can think of this as a long call option on the frontier that is deep in the money. Perhaps this is Google recognizing and saying “We can do Many things, but we can’t do Everything.” ¯\_(ツ)_/¯

[I applaud Google for rationalizing their cash burn. As a former CFO, when I imagine OpenAI, Anthropic, Databricks, etc burning cash year after year and raising equity to cover the deficit, I throw up a little in my mouth 🤢. I wrote about how Nobody hands out metals🥇 for managing cash BUT IT IS CRITICAL in my 10/2025 post, link here.]

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

The Irreversible Ratchet of Technology

Once a paradigm-shifting technology achieves true Product-Market Fit—whether through autonomous LLM agents or foundational AI—there is no historical rewind button. Technology functions as a ratchet: it turns in one direction, locks into place, and makes the prior baseline obsolete. The past is an echo; strategy belongs entirely to the present and the near future.

Consider how rapidly operational baselines shift. “We used to need ten people for this,” “It used to take six hours,” or “This required an entire legacy team”—these statements quickly turn into historical artifacts. We do not revert to inefficient methods once productivity gains are realized. Instead, human capital reorganizes. Labor shifts to higher-leverage tasks, workflows adapt, and in some cases, workers reclaim time—trading five hours of manual grind each week for leisure or strategic focus. Have an emotionally nostalgic moment recognizing this and then (like it or not) move on. The net result is not the end of work, but its structural evolution.

This structural shift directly informs the current debate surrounding capital expenditure and infrastructure buildout. Concerns about whether too many datacenters are coming online in 2026 or 2027 miss the fundamental mechanics of technological adoption. Short-term overcapacity is a common feature of major infrastructure cycles, but markets are dynamic:

• Supply Absorption: We consistently grow into processing capacity as model capabilities and agentic workloads expand.

• Market Clearing: Volume and price will interact to clear the market. As unit compute costs decline, utilization rates increase, unlocking new economic use cases that were previously cost-prohibitive.

• Hyperscaler ROI: The primary cloud providers—Google, AWS, and Microsoft—are positioning themselves as the tollbooths of this new utility. The long-term return on investment is supported by structural demand.

The fundamental question is no longer whether this shift is real or sustainable. We are past the existential debate of how we use to do work. The past is gone. From here on, the discussion is simply about capacity, price discovery, and the rate of absorption.

[Speaking of technology, here is a link to an excellent video of how a natural gas turbine produces an electric current that a data center can use to produce on-site electricity. From my 7/2025 post, Link here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , , , , SK Hynix, TEL

• Housing /Banking - , , ,

• Others- ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Video

Benedict Evans, one of tech’s most widely-read analysts, joins Jacob Effron. The conversation centers on Benedict’s core thesis that comparing AI’s scale to past platform shifts (the internet, mobile, PCs) is analytically useless, and that the more productive move is studying how those previous technologies actually evolved economically to reason about where AI’s value will accrue.

“If you say digital transformation three times then a partner from Accenture will appear in a puff of smoke. “ 😶‍🌫️

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The King of Oil: The Secret Lives of Marc Rich by Daniel Ammann Biography, Business. Great book.

• Streetwise: Getting to and Through Goldman Sachs by Lloyd Blankfein Autobiography. Better than what I was expecting, very good guy.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏-Now we’re just negotiating the price🤑

Regarding AI bubbles, Technology Ratchets, humans being disrupted/slow to change, and market systems helping to find more of an equilibrium. All of this reminds me of a story I have filed in my mind under “now we're just negotiating the price”-

The Louisiana Purchase from Napoleon

In 1802, Thomas Jefferson sent James Monroe and Robert Livingston to Paris with a strict, limited objective: negotiate the purchase of the port of New Orleans and surrounding coastal areas for up to $10 million. At the time, France’s vast territory west of the Mississippi (Louisiana) was considered non-negotiable—it was the crown jewel of Napoleon’s vision for a restored French empire in the Americas. (See the above map.)

The Story:

When the American negotiators first arrived, French officials repeatedly said no, no. They insisted that the French territory was vital to the empire, integral to French ambition, and simply not on the market. Because a “no” is often a “yes” waiting to happen, the Americans decided to be patient.😎

Then, a combination of events hit Napoleon simultaneously: a failed military campaign in Haiti, an impending war with Britain, and an urgent need for liquid cash. Almost overnight, French Treasury Minister François Barbé-Marbois approached the stunned Americans with a completely unexpected counter-offer.

Instead of selling just the small port of New Orleans, France offered to sell the entire 828,000-square-mile territory. The conversation went from “New Orleans is barely for sale” to “How much cash do you have right now💵💰in your pockets?” in a matter of days. The US bought the entire territory for $15 million—pennies on the acre—proving that even continental empires have an instant cash-out price when cash flow tightens.

There is another story of Winston Churchill’s sharp wit while negotiating a clearing price that I find funny, but it is also off-color😳. Link here 😜.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about 🤔

Anthropic and OpenAI Might be Un-Investable

I know everyone is as excited as a puppy 🐶 with two tails for the coming Anthropic and OpenAI IPOs, but I have some serious concerns. Before anyone flashes a term sheet or buys into the pre-IPO retail hype, we need to take a cold, hard look at the structural reality of these businesses.

When you strip away the tech-evangelism, the fundamentals and the geopolitical landscape look incredibly messy for a public market investor.

1. The Valuation: Priced for Perfection (and Then Some)

Let’s look at the sheer weight of the capital structures here. We are no longer talking about nimble startups; these are mega-cap valuations being pushed into the market before the ink on their private rounds is even dry.

• OpenAI: Fresh off its massive $122 billion funding round in March, OpenAI is sitting at an eye-watering $852 billion valuation. While their annualized revenue has scaled impressively to roughly $25 billion, they are trading at a steep 34x annualized revenue run-rate—all while bleeding cash on astronomical compute costs.

• Anthropic: It gets wilder. Anthropic’s $65 billion Series H round in May launched its valuation to $965 billion, briefly eclipsing OpenAI. They are touting a $47 billion revenue run-rate, heavily driven by enterprise coding tools.

Paying 20x to 35x revenue for capital-intensive companies that are essentially sophisticated resellers of compute is a massive gamble. The public markets are being asked to buy at the absolute top of a private funding rocket🚀, leaving virtually zero margin for safety.

2. Open-Source Alternatives: The Structural Price Deflator

In traditional software, you build a moat and enjoy pricing power. In frontier AI, the moat is constantly being drained by open-source (or more accurately, open-weight) alternatives.

Why should an enterprise commit to massive, multi-year seat licenses with OpenAI or Anthropic when Meta is treating frontier-level AI as a loss-leader? Meta’s Llama ecosystem and other open-source models are closing the capability gap at a fraction of the cost. Enterprises can self-host, fine-tune on their own secure hardware, and avoid the vendor lock-in or API dependencies of closed models. When your primary product can be roughly replicated for free by the broader ecosystem, your long-term margins are under existential threat.

3. The Incumbent Squeeze: Competition from GOOG, META, and MSFT

Anthropic and OpenAI are locked in a cage match with the very tech titans they rely on for survival.🔪🩸

• Google (DeepMind) has a vertical integration advantage that neither startup can touch—they own the data, the consumer distribution (Android/Search), and their own custom TPU hardware.

• Meta is weaponizing open-source to destroy the monetization models of standalone labs.

• Microsoft, while a major backer of OpenAI, is a frenemy at best. They hold critical technology rights to OpenAI’s models until 2032, but they are also building out their own internal AI capabilities and infrastructure hedges.

Anthropic and OpenAI are stuck in the middle—forced to buy chips from Nvidia and cloud compute from Microsoft, Google, or Amazon, meaning their capital expenditure essentially flows directly into the pockets of their primary competitors.

4. Federal Government Involvement: The Era of “Gated AI”

This is the ultimate deal-breaker for me. If you think these companies are going to be allowed to operate like normal Silicon Valley growth stocks, you haven’t been paying attention to Washington. We have officially entered the era of government-managed AI releases, and it is a logistical nightmare for commercial predictability.

Look at what happened in June: The U.S. Commerce Department forced Anthropic into a bruising two-week system blackout, ordering them to disable access to their flagship Fable 5 and Mythos 5 models globally for foreign nationals due to national security concerns. Simultaneously, OpenAI had to roll out its GPT-5.6 Sol preview under a strict, opaque “customer-by-customer” government approval process.

National security is now actively slowing down model releases and dictating who these companies can sell to. It gets even weirder on the corporate structure side:

• OpenAI has actively entered early discussions with the administration regarding a proposal to hand over a 5% equity stake to the U.S. government—a move valued at over $42 billion—just to appease political scrutiny and align with national interests.

When a company has to voluntarily surrender billions in equity to the federal government and clear its customer list through the Department of Commerce just to launch a product preview, it isn’t an agile tech disruptor anymore. It’s a heavily regulated, de-facto defense contractor.

The Bottom Line

When these IPOs hit, the media frenzy will be deafening🔊📢. But between the trillion-dollar valuations, the commoditization threat from open source, the absolute reliance on hyper-scale competitors for compute, and a federal government that has its hand firmly on the kill-switch, the risk-reward ratio is completely out of whack.

I’ll be watching from the sidelines. Mark me down as skeptical.

[I don’t know if you heard, but SpaceX went public. 🤭I wrote about Google’s investment size and history in my post 5/2025, here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Housing / Banking - , ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

It is ALWAYS good to remember the chart below because of the intra-year declines.

Podcast

Steve Eisman sits down with Torsten Slok, Chief Economist at Apollo, to break down just how dependent the U.S. economy has become on AI. Torsten walks through the K-shaped economy in detail, explaining why the top third of American consumers are doing fine while the bottom two-thirds are under severe and growing stress. They also discuss the Fed’s impossible position, why Kevin Warsh is unlikely to cut rates anytime soon, and what a slowdown in AI spending would mean for an economy that has become singularly dependent on it. 🌟👍

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• When Everyone Knows That Everyone Knows...: Common Knowledge and the Mysteries of Money, Power, and Everyday Life by Steven Pinker Nonfiction, Psychology.

• The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence by Sebastian Mallaby AI, Biography

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- How Long Would It Take for Waymo to Match 30% of Uber’s Weekly Rides?

I have been thinking about Uber’s moat and how large of a threat Waymo might be.🤔 Below is a direct comparison of both of their operations in the United States as of mid-2026. ( , )

I gave some parameters to Google Gemini and asked it to run the math and a couple scenarios for me. The results were interesting enough for me to want to share them with you.

Step 1: The Target Volume

To find the weekly trip target:

• Current Uber US Volume: Approximately 62.5 million rides per week (extrapolated from US market share and global ride metrics).

• Waymo’s 30% Target: $62.5million x 0.30 = 18.75 million rides per week.

• Waymo’s Current Volume (Mid-2026): 500,000 rides per week.

• The Gap: Waymo must increase its weekly ride volume by 37.5 times to hit this target.

Step 2: Fleet Math & Car Utilization

How hard can a Waymo vehicle work? Unlike human drivers, who average roughly 13 to 18 hours per week on Uber, an autonomous vehicle can theoretically run almost continuously, paused only for cleaning, charging, and maintenance.

The Inputs & Constraints:

1. Operating Limit per Day: Let’s set a realistic upper limit of 20 hours per day per car (allowing 4 hours daily for EV DC-fast charging, vacuuming, and depot maintenance).

2. Average Ride Duration: In dense urban markets, the average trip (including pickup “deadhead” time) takes about 24 minutes (0.4 hours).

3. Peak vs. Off-Peak Utilization: Even with a 20-hour operational window, cars will sit idle during low-demand late-night/mid-day hours. Let’s assume an active passenger-carrying utilization rate of 60% during those 20 hours (12 hours of active paid trips per day).

The Math per Car:

Trips per car per day = 12 hours of active trips/0.4 hours per trip = 30 trips per day

Trips per car per week = 30 x 7 = 210 trips per week

Note: This matches the realistic “high efficiency” targets modeled by autonomous fleet operators (the theoretical ceiling is 40 trips/day, but average real-world utilization sits closer to 20–30 trips).

Step 3: Required Fleet Size

Using our highly optimized utilization rate of 210 trips per week per car:

Required Fleet Size = 18,750,000 target weekly rides / 210 weekly rides per car = ~89,285 cars

To handle 30% of Uber’s US volume, Waymo needs an active US fleet of roughly 90,000 fully autonomous vehicles.

Waymo’s active fleet size as of mid-2026 is 3,871 cars. They need to add approximately 86,000 vehicles.

Step 4: How Long Will This Take?

To project a timeline, we have to look at Waymo’s historical scaling speed, manufacturing partnerships, and capital requirements.

Scenario A: The Current Pace (Linear Scale)

Waymo currently adds roughly 300 vehicles per month (~3,600/year) to its fleet.

• Calculation: 86,000 cars needed / 3,600 cars per year = 23.8 years.

• Verdict: Linear scaling at the current pace is too slow; they would not reach the target until 2050.

Scenario B: Advanced Mass Manufacturing (Exponential Scale)

To get there faster, Waymo relies on major manufacturing partnerships (such as Geely’s Zeekr platform, custom built for Waymo under the “Ojai” project name). If Google/Alphabet and its partners treat this like a standard auto-manufacturing ramp-up, they could transition to a high-volume factory line capable of producing 15,000 to 20,000 AVs a year.

• The Capital Constraint: At an estimated cost of $70,000 per specialized AV (including LiDAR, radar, cameras, and redundant compute), building 86,000 cars requires roughly $6 billion in pure manufacturing capital, not including depot construction, charging infrastructure, and localized mapping operations.

Summary Timeline Verdict

If Alphabet aggressively funds the expansion (leveraging their massive capital reserves and external investment rounds like their recent $16B raise), and faces no crippling regulatory setbacks, Waymo could realistically scale to 30% of Uber’s US weekly rides in 6 to 8 years (roughly 2032 to 2034).

The constraint is no longer the software—it is purely the physical speed of automotive manufacturing, fleet logistics, and local regulatory approval.

6 to 8 years is a relatively short timeline, and all that for the low price of $6 billion. Alphabet/Google has $6 billion in their couch 🛋️cushions. 🤔Very interesting indeed.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about- Daniela Amodei

Daniela Amodei is a prominent American tech entrepreneur and the co-founder and President of Anthropic. She is also the sister to Dario Amodei, CEO of Anthropic.

School History & Academic Background

Unlike many of her peers in top-tier AI leadership, Daniela does not come from a computer science or engineering background.

• She graduated from Lowell High School in San Francisco.

• Undergraduate Education: She attended the University of California, Santa Cruz (UCSC) on a classical flute scholarship. She graduated summa cum laude with a Bachelor of Arts, double-majoring on English Literature, Political Science, and Music.

• College Accolades: Her academic and musical talents earned her top honors at UCSC, including winning the 2008 Concerto Competition as a flute soloist and being selected as a winner of the Senior Thesis Colloquium.

Background & Career Path

Born in San Francisco in 1987, Amodei began her professional career far away from Silicon Valley’s tech core:

• Politics and Global Health (2010–2013): She initially worked in international development at the University of Maryland’s IRIS Center. She then moved into political campaigning, serving as a Field Director for Matt Cartwright’s successful congressional campaign in Pennsylvania before briefly working as a congressional staffer and scheduler in Washington, D.C.

• Fintech and Stripe (2013–2018): Making the jump to tech, she joined the financial services company Stripe as an early employee. Over five years, she scaled through the ranks—starting in recruitment before moving into risk management, ultimately serving as a Risk Manager where she led teams cross-functionally with machine learning and data science to dramatically lower fraud rates.

• OpenAI (2018–2020): She joined OpenAI as an Engineering Manager (leading teams in natural language processing and music generation) and eventually served as VP of People and VP of Safety and Policy.

• Anthropic (2021–Present): Citing differences in directional alignment regarding AI safety and commercialization, Daniela, her brother Dario, and five other OpenAI colleagues left to co-found Anthropic in early 2021. As President, she leads the company’s operations, model alignment, and safety policies.

Family

• Brother: Her older brother is Dario Amodei (born 1983). While raised in the same household, they had drastically divergent early interests—Dario was a pure “science kid” who made the U.S. Physics Olympiad team, while Daniela pursued the arts and humanities. Despite this, they share a remarkably close bond and co-lead Anthropic together (Dario as CEO, Daniela as President).

• Spouse & Children: In August 2017, Daniela married Holden Karnofsky, a prominent figure in the “Effective Altruism” community🤔 and the co-founder of Open Philanthropy. Together, they have one son.

• A Legacy of Activism: The Amodei siblings’ intense focus on ethical boundaries and safety is often traced back to family history. In the 1930s, their maternal grandmother famously chained herself to the Italian consulate in Chicago to protest Italy’s invasion of Ethiopia. Family accounts point to this intergenerational lineage of challenging authority and putting one’s body on the line for ethical beliefs as a driving force behind why Daniela and Dario walk a distinct path in AI safety.

The buzz/tea🍵🫖 /gossip in the tech community is Daniela’s status as a top AI executive who doesn’t code. She has frequently spoken about how her background in literature and understanding human systems has been a massive asset. Her career serves as a prominent modern example that as AI automates core engineering, human skills—such as precise communication, empathy, and organizational risk assessment—become the premium skill sets required to guide the technology. It seems to be working for her. ¯\_(ツ)_/¯

[If you want to learn more, I wrote about Dario in my 6/2026 post, link here. ]

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Engineering

The complex engineering involved in demolishing the I-74 Mississippi River Bridge. Demolishing a bridge has got to be easier than constructing it, right? Not so fast. Very well done, and fascinating.

Video

Laffont, who built Coatue from $50 million to nearly $100 billion in assets, shares why he believes today’s AI buildout looks fundamentally different from the 1999 dot-com bubble, why hyperscaler cash flows change the math on AI infrastructure, and why intelligence itself is becoming the fifth utility, alongside water, electricity, gas, and broadband.

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Fear Index by Robert Harris Fiction. Mix of trading markets and AI.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏

Old People🧓🏻🧓🏾👴🏾👴 Definitely didn’t get old by accident. If you hope to live long enough to get old yourself, you should talk to the oldsters in your life and learn something.

Think about it.🤔 (Setting aside our loved ones that pass on young due to accidents and unfortunate disease.) A large number of those who lived recklessly or adopted dangerous habits died earlier in life.

To get old you have to be a survivor (by definition.) What ever that means.

It is like a 200 year old oak tree surrounded by small and medium sized trees, plus some bushes and scrub. That oak tree has been through a lot.

So, ask an old person (Grandparent, Family member, Friend, neighbor)- What are some secrets to living a long life? How did you get to be this age? Why do you think other people have not lived as long as you? What are things that you would encourage people to do to live long, what are things that you would discourage or ask people to avoid?

Then…. be quiet🤫 and listen. Every now and then… ask them to say more. “Interesting, tell me more about that..?”

And when you are all done, give them a hug and say Thank you. 🙏🏻

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Anthropic CEO Dario Amodei

As the co-founder and CEO of Anthropic—the AI safety and research lab behind the massive Claude language models—Amodei (43 yrs old) is widely considered one of the most influential “architects of AI.” Yet, his path to the top of Silicon Valley looks nothing like the standard tech founder playbook.

Early Life: The San Francisco “Science Kid”

Dario Amodei was born in 1983 and raised right in San Francisco, California. He grew up in a bicultural Italian-American and Jewish-American household that heavily valued precise, disciplined work. His father, Riccardo Amodei, worked locally as a leather craftsman, while his mother, Elena Engel, was a project manager for public libraries.

While the historic dot-com boom of the late 1990s was exploding right outside his bedroom window, Amodei completely ignored it. He later recalled that building commercial websites held zero interest for him; he was entirely consumed by a passion for uncovering fundamental scientific truths. He attended San Francisco’s competitive Lowell High School and made the elite USA Physics Olympiad team as a teenager in 2000.

The Academic Journey: Deciphering Real Brains

Amodei’s university path shifted his focus from the laws of the universe to the mechanics of organic intelligence:

• Caltech (California Institute of Technology): Began his undergraduate work in physics.

• Stanford University: Transferred to complete his Bachelor of Science in Physics.

• Princeton University: Earned his PhD in Biophysics/Computational Neuroscience. His doctoral thesis, which won the prestigious Hertz Thesis Prize, focused on the electrophysiology of neural circuits—essentially mapping how biological neurons network and transmit electrical signals.

• Stanford University School of Medicine: Returned to Stanford for a postdoctoral fellowship, applying mass spectrometry to cellular network models and hunting for cancer biomarkers.

This biological background is crucial to understanding him. While most frontier AI leaders come from computer science, Amodei spent his formative years analyzing how real brains process data.

Career Trajectory: In Chronological Order

1. Baidu (Silicon Valley AI Lab): November 2014 – October 2015. Working as a research scientist under AI pioneer Andrew Ng, Amodei helped build Deep Speech 2. Here, he noticed a critical pattern: scaling up data, compute power, and parameters made neural networks predictably and smoothly better. This realization predated the industry’s official “scaling laws.”

2. Google Brain: 2015 – 2016. As a senior research scientist, his focus pivoted toward safety. He co-authored early, foundational papers on how advanced AI systems could unexpectedly malfunction or cause unintended harm if left unguided.

3. OpenAI: 2016 – 2021. Amodei joined OpenAI early on and rapidly rose to become the Vice President of Research. He set the company’s core scientific direction and directly led the teams that built GPT-2 and GPT-3. He also co-invented Reinforcement Learning from Human Feedback (RLHF), a core technique used to make AI align with human values.

4. Anthropic: 2021 – Present. Deeply concerned that OpenAI was pivoting too quickly toward commercialization at the expense of safety protocols, Dario and a core group of researchers walked out. He co-founded Anthropic as a Public Benefit Corporation, where he serves as CEO.

[Rumored Story: In late 2023, during the chaotic weekend when OpenAI’s board temporarily fired Sam Altman, the board secretly approached Dario Amodei, offering him the CEO position at OpenAI and proposing a merger of the two companies. Valuing his autonomy and Anthropic’s independent safety mission, Amodei flatly declined both offers.]

Family and Private Life

Dario’s most famous family dynamic is also his primary business partnership. His younger sister, Daniela Amodei, is the co-founder and President of Anthropic. The siblings split the empire perfectly: Dario runs the massive compute clusters, machine learning architecture, and technical safety, while Daniela acts as the operational engine—managing multi-billion-dollar cloud infrastructure deals with Amazon and Google, handling policy with Washington, and scaling the company’s business pipeline.

Dario lives a highly private, low-profile lifestyle in the San Francisco Bay Area, near Anthropic’s headquarters. No spouse, partner or children are mentioned. Despite a massive net worth driven by Anthropic’s multi-billion dollar valuations, he intentionally avoids the spotlight, maintains a minimal social media presence, and stays fiercely protective of his privacy.

Amodei doesn’t have the flashy hobbies of a typical tech billionaire. His personal time is dominated by deep intellectual pursuits: reading dense philosophy, studying advanced mathematics, consuming AI safety literature, and writing long-form thought pieces.

To understand Dario Amodei is to understand Interpretability Research. Because of his background as a neuroscientist, he treats AI models not just as software programs to code, but as complex, opaque organisms to dissect. Anthropic spends an immense percentage of its budget looking inside neural networks to map out precisely why an AI says what it says. For Amodei, AI isn’t a race to cross a finish line first—it’s a high-wire act where moving too fast without a safety net could prove catastrophic.

[If you like biographies, I wrote about Alphabet (Google)’s CFO Anat Ashkenazi, in my 8/2025 post, link here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Housing / Banking -

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast

Investor Bill Ackman discussing his evolution in strategy and market outlook. Including- Investment Philosophy, AI Outlook, Activism, Founder-Led Value, Long-term Compounding.

Video

Steve Eisman sits down with Stacy Rasgon, semiconductor analyst at Bernstein, to break down the AI semiconductor boom that has sent the sector up 60% year to date. They explain who is winning, who is playing catch-up, and they close with the most important question in markets right now: what would actually have to happen to derail this boom?

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• American Kleptocracy: How the U.S. Created the World’s Greatest Money Laundering Scheme in History by Casey Michel Nonfiction

• In Praise of Slowness: Challenging the Cult of Speed by by Carl Honoré

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- Take a bit more RISK!

We live in a world that has become obsessively risk-averse. Forget about the ultra rich 1%, I am talking about my neighbors, friends and community. From conservative traditional school curriculums to societal pressures that demand a visible, predictable life path, we are constantly told to play it safe. Even our teachers/professors—bless their hearts, they have a tough job—are structurally incentivized to minimize risk at every turn, trading experimentation for standardized outcomes, and a steady paycheck. But a life entirely free of risk is also a life devoid of truly spectacular creation. As author Neil Simon brilliantly put it:

“If no one ever took risks - Michelangelo would have painted the Sistine floor.”

If you only look down to make sure your footing is safe, you will never look up to paint something beautiful. Accept some more risk to discover what is possible.

To all my friends in their twenties and thirties: this is your permission slip to make a mess. These years shouldn’t be about executing a perfectly sterilized, risk-free plan; they should be about testing your limits, trying the unconventional path, and occasionally skinning your knees. Don’t YOLO, I don’t YOLO with my career or money. Think long and hard before you take action. But, don’t let a safety-first society scare you into sitting on the sidelines of your own career. I accepted thoughtful risk in most jobs, people, investments and career trajectory. Though there were minor short-term setbacks, the net effect was a major compounding of my returns (including fun.)

Apply for the job you aren’t fully qualified for, work in a new location, invest in higher risk for longer term growth. Look up from the floor, embrace a little healthy uncertainty. It will be difficult, but worth it.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

US Homebuilder Consolidation 🏠💰

The U.S. homebuilding industry is in the middle of an unprecedented wave of consolidation. A race for developed land, a industry-wide push toward “land-light” operational efficiency, and massive influxes of institutional capital are reshaping the top tiers of the Builder 100 over the last two years (spanning mid-2024 to mid-2026).

The largest U.S. homebuilding acquisitions during this period highlight this transformation:

But wait, there is more- Dream Finders Homes (a Jacksonville, Florida-based homebuilder) is trying to acquire rival company Beazer Homes (Atlanta, Georgia.) To date, Dream Finders has made four distinct all-cash buyout offers to Beazer's board of directors, 🚫all of which have been rejected: February 5, 2026: $28.50 per share. March 17, 2026: $29.00 per share. May 5, 2026: $25.75 per share. Mid-May 2026: $25.75 per share. The offers were adjusted downward in May to reflect declines in Beazer's stock price and financial performance.

Three major macroeconomic trends are driving these multi-billion-dollar plays:

• The Japanese Capital Surge: Faced with a shrinking domestic population at home, massive Japanese conglomerates like Sumitomo Forestry and Sekisui House are aggressively deploying low-cost capital to capture U.S. market share. Sekisui House recently unified MDC (Richmond American Homes), Chesmar, and Woodside under a single U.S. operating umbrella, while Sumitomo’s acquisition of Tri Pointe pushes them closer to their goal of 23,000 annual U.S. deliveries by 2030.

• Berkshire’s Mega Site-Built Move: Berkshire Hathaway’s massive bet on Taylor Morrison represents its first multi-billion-dollar deal under new CEO Greg Abel. While Berkshire has long dominated manufactured housing via Clayton Homes, this acquisition signals an explicit long-term strategy to build and unify a massive traditional, site-built platform.

• The Land-Light Blueprint: Lennar’s acquisition of Rausch Coleman perfectly illustrates modern structural engineering in housing. Lennar absorbed the asset-light homebuilding operations (the home manufacturing engine), while its spun-off entity, Millrose Properties, simultaneously acquired Rausch Coleman’s physical land assets. This allows the builder to maintain exclusive options on the lots without carrying heavy, illiquid land costs on its core balance sheet.

What else is influencing this consolidation trend?

• While the U.S. remains structurally undersupplied by millions of homes, the operational environment has become incredibly hostile for all but the largest capitalized players. Giants like Lennar, D.R. Horton, and Pulte have internal mortgage subsidiaries. They can afford to buy down a buyer’s interest rate from 6.5% to 4.9% for the entire life of the loan.

• Because large public builders have a fundamentally lower cost of capital, a dollar of profit is worth more to them than to a private regional builder. This creates an arbitrage opportunity where public builders can pay a premium to acquire private operators and instantly make the acquired assets more profitable by refinancing them under public debt structures.

• Building material costs and labor shortages remain major challenges. Mega-builders wield unmatched purchasing power. By acquiring a regional footprint, a national builder can instantly overlay its national supply chain onto the local operation, wiping out ~10% of the localized material costs overnight.

I expect the consolidation trend will continue.

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- ,

• Housing / Banking -

• Others-

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Can you understand why NVDA has a lower PE than WMT and COST? Me neither.

Podcast

Dwarkesh Patel joins Ron Minsky and Dan Pontecorvo at a Jane Street Texas data center for a conversation on GPUs, trading horizons, and hiring.

Video-Why Don’t Computers Just Use One Type of Memory? 🖥️💿🛠️

Your computer has over a dozen different memory and storage locations—from CPU registers to DRAM, SSDs and hard drives—and it needs every single one. In this video, you’ll see exactly why.

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Dark Wire: The Incredible True Story of the Largest Sting Operation Ever by Joseph Cox Nonfiction, True Crime.

• If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All by

  Eliezer Yudkowsky Nonfiction, AI, Science

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏

In the spirit of Memorial Day I wanted to share US Marine Corps rules for gunfighting.🪖 There are some life lessons in there that can also be applied to investing (2,3,4,9,12,13, maybe more).

[PS- Did you know Cadets at West Point Are Trained To Typically Provide Only 4 Responses? I wrote about it in my 1/2025 post here.] Enjoy 🫡

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Impressions of the Berkshire Meeting and Greg Abel (Video down below)

Overall- I would tell you if I thought BRK was headed to hell wearing gasoline pants👖⛽🔥…. It is not. Quite the opposite. I think Greg is bringing the focus, energy and execution that has been absent for perhaps the last 10 to 15 years. Warren was a capital allocator and builder. Charlie was an architect of Berkshire. Greg is a world class operator.

Leaders- BRK is in remarkable hands with Greg. He was joined on stage by Ajit Jain (Vice Chairman of Insurance Operations), and top deputies Adam Johnson (CEO of NetJets and president of Berkshire’s consumer, service, and retail businesses) and Katie Farmer (CEO of BNSF Railway). All of them are impressive and will improve the results of their companies.

Simplification- Greg (and team) is cleaning up complexity at BRK. They are simplifying reporting structure, holding leaders accountable to higher standards, and returning to the clear objective of high measurable returns.

Grandpa Warren- Up front, Warren and Charlie are 2 of my heros. In a world filled with disappointing people, these 2 people are wonderful examples of how to live your life. Respect ✊🏻! Warren remained as CEO too long. He was not bringing the discipline that BRK required, and he is ultimately responsible for the poor performance of the last series of years. I don’t know if it was ego, or like an athlete that loves the game so much they don’t know when to quit, but he should have stepped down sooner. The great investor Phil Fisher also fell into the same trap in his later years, resulting in significant financial underperformance and loss of net worth. We are all human. ¯\_(ツ)_/¯

Greg Abel- I was surprised at Greg’s memory for statistics, trends, regulations in BRK’s wide breadth of businesses. He truly has a command of the drivers and the important facts. Greg is also likeable with a fully formed personality. Greg shared the stage with his management team well. I hope Greg doesn’t talk AS MUCH in future meetings. Now that I am aware of his impressive capabilities, I would rather he answer a question in a concise manner and then move on to the next question. Most people deliver a strong answer, then run it into the ground by rambling (restating their point.) (Greg was guilty of this.) Worse- After we are done saying all the smart things we know, the only thing left to come out of our mouths is stupid stuff. (Greg did not do this.) When we are done with the answer, zip it up 🤐 and be quiet. In the past, Warren answered questions well. Charlie answered questions the best!👍🏻

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , , ,

• Others- , , , $BRK

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast- Anthropic’s CFO

CFO Krishna reveals what it takes to manage the lifeblood of generative AI: compute. From procuring gigawatts of processing power across Amazon, Google, and Nvidia, to dynamically allocating it for model training and customer demand, compute dictates Anthropic’s future. They explore the company’s mind-bending exponential revenue growth, the “cone of uncertainty” in forecasting, and why frontier AI models continue to defy scaling limits. Krishna also shares how Claude is revolutionizing his own internal finance team, Anthropic’s unique culture of safety and transparency, and his bold predictions for AI-driven breakthroughs in healthcare.

[Are you a CFO fan like me? I wrote about an interview of Meta CFO Susan Li in my 6/2025 post here. She is brilliant.]

Video

Berkshire Hathaway CEO Greg Abel presides over the 2026 Berkshire Hathaway annual meeting.

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Inventing the Renaissance: The Myth of a Golden Age by Ada Palmer History, Nonfiction

• Supremacy: AI, ChatGPT, and the Race That Will Change the World by Parmy Olson Nonfiction, Biography, AI

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏-Does M&A Need to be a Circus 🎪?

Over my career, both first hand as a finance professional, and indirectly as an investor, most M&A transactions are a large circus. I think the two most important factors should be- 1. Are the people you are transacting with (your counterparty) honest and examples of high integrity? 2. Is the price fair to you? If either one of this is a no, pack up your bag and head for the door. Any additional time spent is a waste of time. If both of these are true, then a larger circus is unnecessary and a waste of funds.

Munger often recounted a situation where Berkshire was involved in a legal proceeding or a due diligence inquiry related to an acquisition. When investigators or opposing counsel demanded to see the “staff papers”—the thick binders of projections, synergy reports, and consultant studies typically generated by big corporations during M&A—Munger’s response was:

“Not only do we not have any staff papers, we don’t have any staff!”

Why? If Buffett and Munger answered yes, to the questions at the top, then they got on the phone with each other for an hour to work out the details. That was it. Now, both of them were masters at business, and Buffett had an encyclopedic knowledge and memory of their target company and industry. But, no additional study or input was necessary.

Chapter 6 in The Outsiders, outstanding book, is about Bill Stiritz unbelievable compounding as CEO of Purina. The author writes-

He was well known for showing up alone to important due diligence meetings or negotiations where the other side of the table was crowded with bankers and lawyers. Stiritz relished the unorthodox approach. A then junior banker at Goldman told me of a late-night due diligence session during the RJR Nabisco sale process when Stiritz came to a conference room at the Goldman HQ alone👨🏻‍💼, armed only with a yellow legal pad, and proceeded to walk through the key operating assumptions one by one before making a final bid and going to bed.🎤🫳🏻🛌🏻

Look for companies who’s leaders behave like these legends. Be cautious when companies have high drama, long and messy transactions.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

MAG Tech ownership of Impending IPOs-

As near as I can tell, below is a summary of how much and what percentage each Magnificent Tech company owns of each large tech IPO.

I don’t like to repeat all the same media statements of this group, so here are a few other observations:

• Big Tech owns 37% of OpenAI and 35% of Anthropic. 🤔Interesting.

• MSFT, AMZN and GOOG all own percentages that add up to about 20-28% across the 3 IPO candidates.

• Google is the only investor in SpaceX.

As I have aged and gotten more experience, I have grown an opinion that is generally UN-favorable of IPOs. They are priced too high, the share allocation is hard to get, trading on the first day is frantic, the financials are window dressed, and I could go on. I am pretty confident that I will not be a participant in any of the 3 potential IPOs listed above…. HOWEVER, I do own all of the big tech investors in the table above and therefore, I do also already own the 3 IPO candidates. I honestly much prefer this method of participation to the YOLO experience of the IPO day.

IPOs remind me of a residential real estate investor friend of mine. He owned a bunch of properties and from time to time he would sell one. I was talking to him 20 years ago about buying and selling properties and he shared with me a wonderful nugget-

You don’t make your money when you sell a property. Nick, you make money when you Buy the property. Either you get an amazing deal up front, or you are 90% certain that the appreciation will be significant over the coming 2-3 years. By the time you sell it, all the money has been made. Selling is about minimizing the expenses you incur to sell. It is always better to be a seller than a buyer.🎤🫳🏻

It is always better to be a seller than a buyer.🤔

[If you are interested in how Google owns ~6% of SpaceX, I wrote about it in my 5/2025 post here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , ,

• Housing / Banking - , ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

In a historic move, the United Arab Emirates said it will leave OPEC next month, a significant blow to the group. UAE was the 3rd largest producer.

Switzerland is 100% electric powered trains, on the right is the USA 1%. 🤦🏻‍♂️

Podcast- Google’s Liz Reid on Who Will Own Search in a World of AI

Elizabeth Reid is head of search at Google. Liz has been with the company for more than two decades, witnessing multiple tech transformations in that time. They talk with her about how Google is incorporating Gemini into search via AI overviews, what that means for traffic and sales, and the practical experience of search in an age of LLMs and internet slop. (I got a lot more signal from it than I would have expected.)

Video-Mental Models for Running Businesses

Mohnish Pabrai is a great communicator and a lot of fun. You might remember that he is a buddy of Charlie Munger. This talk was given 3/16/2026, 45 minutes.

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Tracers in the Dark: The Global Hunt for the Crime Lords of Cryptocurrency by Andy Greenberg Nonfiction, Crime, Technology

• The Power of the Dog by Don Winslow Crime Fiction.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- Meta 30-Day Hunger Games

I have heard the media repeating 2 things about Meta recently. They are laying off 10% (8,000 employees) of their workforce on 5/20/26, and second, they are Immediately installing tattleware on all employees computers. Interesting🤔… Why?

The move to install granular activity tracking — often called “tattleware” — is a high-pressure tactic that signals a shift from Meta’s historical “we are all responsible adults here” culture to a “Metrics” culture. While the official reason is often security or resource optimization, the psychological and operational impact on a workforce facing a 30-day layoff window is intense.

1. The Performance Floor Strategy

This tool isn’t necessarily designed to find the highest performers; it’s designed to identify the Bottom 10-20%.

• Automated Auditing: In a company as large as Meta, managers often don’t have eyes on every sub-contractor or junior engineer. Tracking clicks, active window time, and peripheral usage (mouse/monitor) provides an automated low-activity report.

• The Cut-Off: If Meta needs to cut 8,000 people and has 5,000 low-activity pings, the decision making process for HR becomes significantly faster and more defensible against potential legal pushback.

2. Forcing The Hunger Games (30-Day Sprint)

Employees know they have 30 days and this forces a “pick up the pace” mentality, but it often produces Quantity over Quality:

• Performative Productivity: When employees know their clicks are being tracked, they tend to stay in active windows longer and avoid deep long thinking time that doesn’t register as activity.

• The Anxiety Loop: For workers in the 30-day window, this creates a treadmill effect. They may work 12-hour days just to ensure their activity dashboard looks green, even if that work is less impactful than a few hours of focused coding or strategy.

3. Why now? (The Timing Factor)

Deploying this exactly when a layoff date is set serves a specific psychological purpose: Managing the Lame Duck period.

• When some people know they might be fired in 30 days, they stop working entirely and spend all day on LinkedIn.

• By installing tracking tools now, Meta is essentially saying: “We are still watching you. If you go ‘dark’ during this month, you are guaranteed to be on the May 20 list.”

MY SUMMARY: This is a Meta/Hunger Games 30-Day survival environment. It reminds me of a philosophy from Netflix Founder and former CEO Reed Hastings-

We’re a TEAM, not a family.

A family is about unconditional love, despite your siblings’ bad behavior. A high-performance team is about hiring, developing, and cutting smartly, so we have stars in every position.

Metrics at Meta are going to prove that employees are locked-in and shipping code, or managers are going to identify them for dismissal.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about- Nvidia’s Financials

TSMC and ASML released their respective quarterly financials this week, and both were excellent. With any company that I follow or am interested in, I read the press release and work through the financial statements. On many companies, I also prepare a 10+ year DCF. After I have done this [fetch and brain preload 🧠] work, I listen to the earnings call.

As TSMC and ASML are doing so well and have strong guides for the rest of calendar year 2026, I started to wonder how Nvidia might stack up.

ASML sells → TSMC sells → NVDA sells → GOOG & MSFT. Below is how I wanted to dimension the group for myself.

I think markets are often rational but not always. It is like a neighborhood bar that has a rational owner operator running the bar selling beers for $4. But from time to time the owner steps out for an hour to grab a hair cut, and leaves his drunken friend to mind the bar in his absence…🫤🤪🥴 beers might sell at $2 for a little while. 🍺🍻

Below are some questions I am not going to answer, but I am going to let us all turn them over in our minds-

• Why does ASML have a PE of 55 and NVDA has a 41? Does that make sense?

• Does MSFT deserve the lowest PE at 26, of the group? It has some of the highest gross and net margins.

• Who has the highest revenue growth potential in group, or how would you rank them?

• The net margins for all of them are yummy, right. 😋😉

• Does NVDA or TSMC carry the most risk? TSMC has geographic risk, but NVDA buys all of their chips from TSMC, soo…..

Think about it, and talk amongst yourself. 👍🏻 [ , , , , ]

[If you are interested in ASML, I wrote about how China was having trouble replicating their chipmaking equipment on an 11/25 post, link here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Housing / Banking - , , ,

• Others-

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

If you enjoy my charts I posted a bunch of them on x this week, check them out here, x.com/NickeNumbers

Podcast- OpenAI’s Identity Crisis, Datacenter Wars, Market Up on Iran News, Mamdani’s First Tax

This was a very good episode. Also, see my Wild Card section below as I write about a segment of the show for us to focus on.

Video- The Logistics of Oil Near Iran (17 Min)

Explores how the shutdown of the Strait of Hormuz during the current Iran conflict impacted global oil prices despite a relatively modest supply shock. While 20% of the world’s oil historically traveled through this choke point, global production only dropped by about 8%, yet prices spiked by 60%.

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Atomic Accidents: A History of Nuclear Meltdowns and Disasters: From the Ozark Mountains to Fukushima by James Mahaffey History, Science.

• Cool: How Air Conditioning Changed Everything by Salvatore Basile History, Technology.

• Confessions of the Pricing Man: How Price Affects Everything by Hermann Simon Business, Economics

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- OpenAI vs Anthropic (Capital vs Growth)

In the recent All-In podcast that I reference above in the podcast section, also linked here, the group talks about OpenAI raising Capital contrasted against Anthropic’s incredible Revenue growth. The discussion was so thoughtful I wanted to call it out specifically.

The time mark references start and end around here- 15:23 to 20:16

Q- …there’s a big debate going on amongst the investors. The Financial Times piece also questioned the $850 billion valuation of OpenAI. Secondary markets have now priced Anthropic higher than OpenAI for the first time. This is the flipping that people predicted. One investor said OpenAI would need to IPO at a valuation of 1.2 trillion for the last round to make any sense, but there’s no buyers currently at the $850 billion valuation that OpenAI just closed.😬 According to Bloomberg, Travis, how do you handicap this race between these two leading frontier models?

A- Growth is 👑king right now in this world, in this segment. Growth is the whole damn thing. And if Anthropic is growing faster than OpenAI by a significant clip, the investors right now are going to play it forward. And they’re it, you know, you start to get network effects around compute, network effects around the number of tokens you’re pushing out for various customers, enterprise or consumer. And ultimately, you know, it’s not great today, but how that plays into reinforcement learning and the things getting smarter over time. There’s so much upside to volume and scale that if they are growing faster at the same size, even if OpenAI is still growing, but if they are half the growth rate, a third the growth rate, a fifth the growth rate, I would be I’d be worried.😨

Q- And you saw this at Uber specifically when you accelerated away from the competitors like Lyft and Door Dash. Yeah.

A- Yeah. Network effects was the whole thing at the end of the day and network effects was based on scale. So yes, I’m sort of like coming from my very specific experience. But if you believe there’s network effects with the scale of data that you have and the scale of customers and the revenue that this is cash that’s coming in that you redeploy into compute and so say there’s network effects of large compute I’d be very worried uh if I’m OpenAI and seeing somebody growing faster at the same size.

…

🚨you can use capital and investment to acquire scale and network effects associated with it. But if somebody is getting that scale with revenue and let’s call it contribution margin, contribution profit, efficiency will outstrip subsidy and you can’t just keep raising hundred billion dollar things forever. That’s where the train will stop. And if Anthropic is funding theirs through revenue and other folks [OpenAI] are funding it through investment, that’s a short-term solve, but it the long run is whoever is scaling their actual usage and system and ultimately with contribution profit that then soaks up the need for investment. That’s a very scary machine if you’re competing against it.🚨

Part of what the guys are nibbling on is that Anthropic appears to be compounding their Revenue growth faster than OpenAI. And if you have the momentum of exponential growth behind you, it can be like a ride on a rocket ship🚀. No amount of OpenAI capital raise will be able to make up for the Anthropic revenue growth difference.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about- Fast Growing TEXAS

Population trends in the United States have shifted significantly over the last decade, particularly following the 2020 Census.

Fastest Growing- The South and West continue to dominate growth, largely driven by domestic migration from the Northeast and Midwest.

State 2020 Population (Base) 2025 Population Total Growth

Texas 29,145,505 31,709,821 +2,564,316

Florida 21,538,187 23,462,518 +1,924,331

10-Year Trends (2016–2026)- Looking at the full decade, the disparity between “Sun Belt” states and the “Rust Belt” or Northeast becomes even more pronounced. Texas & Florida alone accounted for nearly 40% of all U.S. population growth over the last decade.

Texas is projected to surpass California in population by 2045.

The largest US states by land size rank order are Alaska, Texas and California. Texas is 269K square miles compared to California’s 164K. This makes Texas roughly 64% larger in total area. The migration to Texas remains one of the most significant demographic shifts in the U.S. as we move through 2026. While the “California-to-Texas” pipeline is a well-known narrative, the drivers have evolved from simple “cheap housing” to a complex interplay of corporate strategy, tax policy, and lifestyle trade-offs.

Corporate Migration: The “Big Tech” and “Industrial” Shift

Companies aren’t just moving for lower costs; they are moving to be closer to new talent hubs and avoid the regulatory friction of the West Coast. More than 300 companies have moved their headquarters to the state in the last nine years alone.

• Recent High-Profile Moves (2024–2026): * Chevron: Relocated its global headquarters from San Ramon, CA, to Houston in late 2024, citing the need for a more business-friendly environment for the energy sector.

  • SpaceX & X (formerly Twitter): Elon Musk moved these headquarters from California to Bastrop and Starbase, Texas, in 2024.

  • Realtor.com: Announced a move from Santa Clara, CA, to Austin in early 2025 to tap into the state’s booming real estate and tech talent pool.

• Infrastructure Magnets: It’s not just HQs. Huge projects like Samsung’s $44 billion semiconductor plant in Taylor and Apple’s $1 billion North Austin campus act as “anchor tenants” that pull thousands of smaller supplier companies and service providers into the state. (Long Texas cement and aggregate supplier . )

• Other noteworthy companies who have relocated to TX- HPE, CAT, Toyota, Schwab, CBRE, XOM, Oracle.

The Role of State Income Tax

The lack of a state personal income tax is a primary mechanical driver for both individuals and high-earning professionals.

• The “Wealth Gap”: In 2026, the contrast is stark. California’s top marginal income tax rate remains the highest in the nation (exceeding 13%), while Washington state has introduced capital gains taxes that have pushed some high-net-worth individuals toward Texas’s zero-tax regime.

• Wealth Migration: Recent IRS data shows that high-tax states like California and New York lost billions in adjusted gross income (AGI) to states like Texas and Florida. For a professional making $200,000, moving from San Jose to Austin can result in an immediate “raise” of roughly $15,000–$20,000 just in tax savings.

[Note on Property Taxes: While Texas has no income tax, it has some of the highest property taxes in the country (often 2%–3% of home value). Many movers find that while they save on their paycheck, their annual tax bill to the county can be a "sticker shock" if they aren't prepared.]

Jobs and Economic Growth

While California still has the largest labor market, its growth has flattened (0.0% growth in some 2025 metrics). Conversely, Texas continues to lead in job creation +0.8%.

Why People Are Leaving CA & WA

It is rarely a single reason, but rather a “push-pull” dynamic where the “push” factors on the West Coast have become unsustainable for many.

• Housing Affordability: The price-to-income ratio in California is roughly 8.2, meaning a home costs more than eight times the median annual income. In many Texas metros, despite recent price increases, that ratio remains closer to 4.0 or 5.0.

• Cost of Living (COL): California’s COL index is roughly 142 (where 100 is the national average), while Washington’s is 114. Texas generally sits between 90 and 95, providing significantly more purchasing power for the same salary.

• Regulatory “Friction”: For entrepreneurs and small business owners, the cost of compliance, permitting, and labor laws in WA and CA has led to “fatigue,” making the Texas “Open for Business” mantra highly attractive.

What Else is Causing the Trend?

Beyond the spreadsheets, there are “soft” factors driving the 2026 migration:

• The “Second-Tier” Boom: People aren’t just moving to Austin anymore. Cities like Fort Worth, San Antonio, and El Paso are seeing surges because they offer the “Texas Advantage” without the skyrocketing traffic and prices now found in Austin.

• Housing Supply: Texas builds more houses. In 2025, Texas issued significantly more building permits than California, despite having a smaller population. This “pro-growth” zoning makes it easier for a middle-class family to actually buy a new home rather than bidding on a 50-year-old bungalow. (Long Texas builder , or another Texas homebuilder .)

• Political & Cultural Alignment: There has been a documented trend of “ideological migration,” where people move to states that better reflect their personal views on governance, education, and pandemic-era policies.

Closing

The Great Migration to Texas is no longer just a post-pandemic anomaly; it is a structural realignment of the American economy. As capital, corporations, and talent continue to cluster in the Sun Belt, the ripple effects will be felt across many asset classes—from residential real estate and commercial infrastructure to the regional banks financing this growth.

While the “Texas Advantage” of zero income tax and pro-growth zoning provides a clear tailwind, savvy investors must also weigh the “friction” of rising property taxes and infrastructure demands. Ultimately, being aware of these demographic shifts isn’t just about tracking where people are moving—it’s about understanding where the next decade of economic value is being created.

Podcast- Rear Admiral (Ret.) Mark Montgomery

Michael speaks with Rear Admiral (Ret.) Mark Montgomery, Senior Fellow at the Foundation for Defense of Democracies, following his recent trip to Ukraine. Ukraine’s Persistence, Baltic ‘Gray Zone’ Attacks & Russia’s Weak Spots. 👍

Video- The China Hu$tle

An unsettling and eye-opening Wall Street horror story about Chinese companies, the American stock market, and the opportunistic greed behind the biggest heist you've never heard of. Also featuring Carson Cutler Block, American investor and activist short-seller (AKA- Muddy Waters Research), and Jim Chanos short-seller. The movie trailer is below, and can be watched on Amazon.

Carson Block (Muddy Waters Research) —“When somebody breaks the law here (USA), there is a process of accountability. At some point in a China-based company, that accountability ends… And it ends at the ocean🌊.”

…. because the U.S. lacks the leverage to force a change, and China lacks the incentive to reform a system that effectively transfers American wealth to its own citizens. ⚠️As an investor, Buyer beware!☠️

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Technological Republic: Hard Power, Soft Belief, and the Future of the West by Alexander C. Karp Nonfiction, Technology, Politics,

• An Engine, Not a Camera: How Financial Models Shape Markets by Donald Angus MacKenzie Finance, Economics, Nonfiction. Nerdy, but if you like the subject it was very good.

• A Curious Mind: The Secret to a Bigger Life by Brian Grazer Nonfiction, Business.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- Endless Action/Reaction

The media generally write for the average reader. That means stories are framed with clear winners and losers, simplified enough to digest in five minutes. They give us a hero, a villain, and two or three tidy takeaways. To any sophisticated investor, it’s no shock that life is rarely that convenient; it is infinitely more nuanced.

Most situations involve a complex web of actors. Look at the news from the last two weeks regarding Meta, Google, Apple, or even the U.S. Navy vs. Iran. There are always multiple parties in play, and each side gets a vote. The media would have us believe that “X will happen” or “Y is the loser.” They feed us binary labels: “Apple is the winner,” or “Meta is bad for teens.” You already know this, but it’s worth saying plainly: it’s nonsense. That’s a narrative for children and fools, and we are neither.

The Evolution of the Play- When the “other side” gets a vote, the game evolves. Each side is fueled by their own creativity and resources. In fact, a lack of capital can be a massive advantage; it forces an opponent to find asymmetric, creative responses that a well-funded incumbent wouldn’t even consider.

Unrelenting Determination- We also need to remember that while the games we watch—chess, football, soccer, basketball—have a clock, real-world conflict is often endless. Some opponents have an unrelenting determination to win, while others are simply tenacious enough not to lose. These are similar, but not the same. A refusal to lose can wear down the expected winner and eventually hollow out the value of their “victory.”

Meta has lots of resources and a determined founder as CEO. Apple has a huge installed base of users and can continue to fail at AI until it finally gets it right. Microsoft is down for now, but they are a fierce competitor. The U.S. Navy and Iran can continue to strike/counterstrike with missiles and words for a very long time. Life is hard, complicated, and the enemy always gets a vote. Wishing it were otherwise only works in Disney🧚🏻 movies.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Low Cost High Volume Token Generation

At GTC Jensen lays out the economics of AI for the larger ecosystem and makes the strong case for Nvidia’s platform-

Would you like to have higher revenues or lower? Would you like to have higher ASPs or lower? Would you like higher margins or lower? Because that’s all it means.

Look, TSMC’s wafers are the highest in the world, but they’re the best value in the world. And I gladly pay for it. ASML systems are the most expensive in the world; they’re worth it. There’s no question about it.

And so the question is simply: do you want to make more money or do you want to buy the lowest cost equipment? Do you want to make more money or do you want to buy the lowest cost equipment? That’s the difference.

Now, what I just said is a new concept and I think we can all acknowledge that. I just treated a computer system the way I treat TSMC’s chip factory, the way I treat ASML manufacturing equipment. And that’s not the way people thought about it in the past.

If I have two CPUs, one of them is 256 cores, the other one’s 256 cores. Tell me which one is the better one? Well, the cheaper one’s the better one because I’m renting it by the core anyways. But that’s not the way tokens are created. You don’t rent by the core. You monetize by the tokens per second, and so it’s different economics.

Does that make sense? You’re not renting cores. You’re not renting nodes. You’re producing tokens, which is the reason why everything changed. It was necessary to make sure that everybody understands the economics of the new world.

Are you Excited😃 about AI, or afraid😨 of an AI future. Do you like Nvidia, or dislike Nvidia. Regardless- Nvidia makes the absolute best token factories, selling at the highest prices… and they will be dominating far into the future. ¯\_(ツ)_/¯

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- at GTC, ,

• Housing / Banking -

• Others- ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast- Ada Palmer on

Why Leonardo was a saboteur, Gutenberg went broke, and Florence was weird.

Renaissance history is so much wilder and weirder than you would have expected. Very fun chatting with Ada Palmer (historian, novelist, and composer based at the University of Chicago). Listening to Ada is like drinking from a fire hose.👍

[I overloaded on learning Italian language electives during my accounting undergrad, and spent a month in Florence after graduation. 😁]

Video- Dylan Patel also on

Dylan Patel, founder of , provides a deep dive into the 3 big bottlenecks to scaling AI compute: logic, memory, and power.

And walks through the economics of labs, hyperscalers, foundries, and fab equipment manufacturers.

[Dwarkesh is on fire 🔥🔥👍 Good for him.]

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• How Infrastructure Works: Inside the Systems That Shape Our World by Deb Chachra Engineering, Nonfiction.

• On Democracy by E.B. White Nonfiction, History.

• Underground Empire: How America Weaponized the World Economy by Henry Farrell Nonfiction, business, politics.

• The State of the Art by Iain M. Banks Science Fiction.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- I don’t care about being right

I have a mixed opinion of the investor Ray Dalio, but something I heard from him has stuck with me for many years and I want to share it with you all. I have found it to come up again and again, and it also illuminates🔦 my thinking and decision making.

Dalio-

“I don’t care about being right; I care about what is right. And I want to know how we know what is right.”

Do you know someone that always has to “BE” right. Have you been in a group where you know the final decision is wrong, but no one is willing to speak up. Have you finally found yourself to be wrong😲, but you were too stubborn to admit it sooner. My answers- Yes, Yes, and HELL YES.

Dalio argues that most people are hindered by two “barriers”: their ego and their blind spots. To overcome these, he suggests that you must stop trying to prove you have the answer and instead focus on the process of finding the best answer, regardless of who it comes from.💡

In his framework, this involves:

• Replacing the Joy of Being Right with the Joy of Learning What’s True: Shifting your emotional reward from “I knew it” to “I found the truth.”

• Triangulating with Believable People: Actively seeking out the smartest people who disagree with you to see if you are missing something.

• Systematizing the “How”: Using data and “believability-weighted” decision-making to determine “how we know” an idea is actually the best one.

Find a process to determine what is right, and work to keep your ego out of the decision. I think Buffett sums this up well for investing- “You are neither right or wrong, if people agree with you. You are right if the facts make you right.”

[As it appears I am handing out advice, tell your kids that → Cadets at West Point Are Trained To Typically Provide Only 4 Responses.. And I wrote about it in my 1/2025 post here.]

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Can Google Gemini Pick Stocks like Warren Buffett?

I devised a challenge for Gemini in the last week that I thought it would fail- Can Google Gemini Pick Stocks like Warren Buffett?

Here is my exact prompt- Can you identify 8 publicly traded companies that meet these requirements: 1. Good businesses, 2. that are obvious to understand why they are good, 3. each company with a durable, competitive advantage, 4. run (managed) by able and honest people, and 4. available at a price that makes sense (reasonable price relative to its fair intrinsic value.)

I am providing you with the results below, but before you dive into it, I am going to give you my quick thoughts and a grade for Gemini.

SELECTION- Gemini did follow my instructions very well, and it selected some excellent companies. Although I am familiar with all of them, I would only call myself an expert in a few of them. For its screening and selection, I am going to give it an A. (Believe it or not.) I think its justification for each companies moat, was impressive.

QUALITATIVE- One of the things that I spend a lot of time on is getting a feel for the managers of my companies. CEO, CFO, etc. Are they honest, hard working, shrewd, smart, talented, experienced, committed, clear thinking. I build this opinion through listening to them talk. In the Management section, Gemini provided information, but it was weak. It is confident about what it wrote, but there is no “signal” in the content. For this reason, I am going to give Gemini a C+.

OVERALL- I am surprised to say, I am pretty impressed. And, I have homework to do because I need to learn more about the companies that Gemini selected that I do not know enough about.

And with that, Drum roll, Please- 🥁🥁🥁🥁🥁🥁🥁- Gemini take it from here-

🤖-Finding companies that satisfy all of Warren Buffett’s four main criteria (quality business, understandable, durable moat, honest management, and fair price) is a high bar. As of late February 2026, the following eight companies are frequently highlighted by analysts and value investors as meeting these requirements.

1. Alphabet (GOOGL)- Alphabet dominates global search and digital advertising through Google and YouTube.

• The Moat: It has a massive network effect and brand power; “Googling” is literally a verb. YouTube remains the dominant video platform for creators.

• The Management: Sundar Pichai and the executive team have a long track record of disciplined growth and massive cash flow generation.

• The Price: Despite its size, it often trades at a reasonable P/E ratio (forward ~22x) relative to its growth, especially given its leading position in AI integration.

2. Visa (V)- Visa is a global payment technology company that connects consumers, businesses, and banks.

• The Moat: A “toll booth” moat. It doesn’t lend money (so no credit risk); it just takes a tiny fee on $16 trillion worth of annual transactions. Replacing its global infrastructure is nearly impossible for a competitor.

• The Management: Known for extreme capital discipline, returning billions to shareholders via buybacks and consistent dividend hikes.

• The Price: As of early 2026, it is viewed as a “Core Stock” available at a fair value compared to its reliable double-digit earnings growth.

3. American Express (AXP)- AmEx provides charge and credit cards, specifically targeting an affluent, high-spending clientele.

• The Moat: A closed-loop network and “prestige” brand. Because AmEx is both the card issuer and the network, it earns more per transaction than peers. Its customer base is more resilient to economic downturns.

• The Management: Management has successfully pivoted to attract younger, high-earning Millennial and Gen Z users, keeping the 175-year-old brand relevant.

• The Price: It remains a top Berkshire Hathaway holding (over 20% of their portfolio) and is often valued reasonably relative to its high return on equity.

4. S&P Global (SPGI)- A leader in financial information and credit ratings.

• The Moat: An oligopoly moat. Along with Moody’s, they control the vast majority of the credit rating market. Companies must pay them for a rating to issue debt to the public.

• The Management: Led by executives who focus on high-margin data services and recurring subscription revenue.

• The Price: While rarely “cheap,” it is currently cited as a wide-moat stock with secular tailwinds from data consumption that justify its current entry price.

5. Coca-Cola (KO)- The world’s largest beverage company with a portfolio of over 500 brands.

• The Moat: Brand equity and the most sophisticated distribution system on Earth. If you want a Coke in a remote village in the Andes, they can get it there.

• The Management: A “Dividend King” with 63 years of consecutive increases, showing deep commitment to long-term shareholder value.

• The Price: Often used as a defensive play; in early 2026, it is considered a “safe stock” with a reliable ~3% yield that protects against market volatility.

6. The TJX Companies (TJX)- The parent company of T.J. Maxx, Marshalls, and HomeGoods.

• The Moat: Inventory expertise. They have a “treasure hunt” shopping experience that is very difficult to replicate online. They buy surplus name-brand goods at massive discounts that others can’t access.

• The Management: Management consistently produces some of the highest Return on Invested Capital (ROIC) in retail (over 45%).

• The Price: It is currently listed among the top quality-value stocks for 2026 due to its consistent ability to beat earnings estimates.

7. Caterpillar (CAT)- The world’s leading manufacturer of construction and mining equipment.

• The Moat: An irreplaceable dealer network. When a $5 million mining machine breaks in the middle of a desert, Caterpillar’s global parts and service network is the only one that can fix it fast.

• The Management: Management is currently diversifying into data center power solutions, showing adaptability to the AI boom.

• The Price: Considered a bellwether of the global economy, it is currently trading at a level that value analysts call “Standard” to “Fair.”

8. Automatic Data Processing (ADP)- A global provider of cloud-based human capital management (payroll) solutions.

• The Moat: High switching costs. Once a company integrates its payroll, taxes, and benefits into ADP, it is incredibly painful and risky to leave.

• The Management: Known for a “Standard” but highly reliable capital allocation strategy that favors long-term stability.

• The Price: Morningstar currently rates ADP as a “Wide Moat” stock trading at a discount to its fair value (Price/Fair Value of ~0.73).

[You notice that Google Gemini’s #1 stock pick was Alphabet? 🤭]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , ,

• Housing / Banking - , , ,

• Others- ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

If you like my graphs and charts, this week I am going to direct you to my account on X. I post a lot of my visuals there, and there were too many for me to choose from. I enjoy them all. If you do also, go check them out x.com/NickeNumbers

Video- Rare Peter Lynch Interview

Josh Brown⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ is joined by Peter Lynch, Vice Chairman of Fidelity Management and Research to discuss: Peter’s legendary career, how individual investors can succeed in the market, the biggest investing mistakes, and much more!

If you don't understand what you own, you're toast. 🔥🍞

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Perfect Weapon: How the Cyber Arms Race Set the World Afire by David E. Sanger Nonfiction, Technology. Sundar Pichai recommendation.

• The Wedding People by Alison Espach Fiction, Best Seller, Relationships.

• Positioning: The Battle for Your Mind by Al Ries Business, Marketing

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- AI Bubble, or Room to Run 🏃🏽‍♀️🏃🏾‍♂️🏃🏿‍♂️🏃🏻

Stop scrolling for a second and look at these numbers.

8.1 billion people on Earth. 6.8 billion have never used AI. Not once. Not ChatGPT, not Claude, not even a voice assistant. Never.

Another 1.3 billion have tried a free chatbot—probably asked it one question, got a response, then never came back.

Only 15-35 million people worldwide actually pay for AI tools. That’s 0.3% of humanity.

And the 2-5 million using AI for coding or development? That microscopic red line on the visualization? That’s us.

The above information was written by Aymane Yousfi for the publication Medium. And it gets me excited. Think about how much productivity is waiting to be unlocked when more and more of humanity has access to AI.

[Side bar, ever wonder how fast a computer can think? I wrote about it in my 6/2024 post, link here. Seconds → Milli → Micro → Nanosecond…⚡]

Other ideas and observations I have-

• We need a lot of infrastructure to serve this population. And it is going to take years, maybe more than a decade to roll it out. ( , , , )

• We are going to need energy to power it all.

• We are going to need hardware ( , , , , , ) ,

• We are going to need computer chips ( , , SK Hynix, , , ) .

• As more humans are getting access☝️, the models are getting more efficient👇, but humans are finding more use cases☝️, but the hardware is also getting more efficient👇.🤔

• And then there are personal agent orchestrators, or just personal AI Assistants that can fire off 10 different searches and sub-agents. When these have wider and wider adoption by the population. 🚀🚀

“As models move to FP4 precision (Blackwell/B200 architecture), they consume significantly less energy per token. However, because agents now run ‘loops’ (a single human prompt triggering 10-20 sub-tasks), the net compute demand is projected to grow by 1000x by 2027, despite the efficiency gains.”

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

The Invisible Assembly Line: Why Large Builders Win and Innovation Stays Slow

If you walked a 1920s surgeon into a modern operating room, they’d be lost. If you walked a 1920s framer onto a modern job site, they’d probably ask where you keep the water cooler🫗 and then get right to work.

While every other industry—from automotive to aerospace—has embraced the efficiency of the assembly line, single-family home construction remains stubbornly “bespoke.” Here is why innovation in building is so difficult, and why the “factory-built” dream often dies on the back of a tractor-trailer.

1. The Logistics of the “Lumbering” Giant

The dream of Prefabricated Housing is often thwarted by the reality of the Interstate Highway System.

When you build a house in a factory, you are limited by the physical dimensions of a standard tractor-trailer: roughly 8.5 feet wide and 13.5 feet high. If you want to ship a pre-built room, you’re either shipping a lot of “expensive air” or you’re cutting the design into awkward, narrow slices.

Furthermore, a house is a rigid structure designed to sit still for 100 years—not to bounce at 70 mph over potholes. The vibration and “road-shock” during transport often lead to hairline cracks in drywall, shifted frames, and window seal failures. By the time a prefab module arrives, the “speed” gained in the factory is often lost to the “repairs” needed on-site. [Some of my thinking on this process is informed by the outstanding work that does on his Substack.]

2. The “Prototype” vs. The “Product”

In manufacturing, the first unit is the most expensive; the 10,000th is the most profitable. Construction rarely gets to unit 10,000.

• Bespoke is the Enemy of Speed: Most homebuyers want to feel unique. They change the floor plan, move a wall, or upgrade the kitchen. This makes every house a “prototype”—a one-off experiment that requires human problem-solving on the fly.

• The Regulatory Patchwork: Even if you built the “perfect” modular home, it might be legal in Dallas but a code violation in Denver. Between varying snow loads, wind speeds (hurricane zones), and local zoning quirks, a truly “national” assembly-line home is a regulatory nightmare.

3. The “De Facto” Assembly Line: Large vs. Small Builders

Innovation in construction today isn’t happening in high-tech factories; it’s happening in process management.

Large-scale builders (the “Whales”- , , ) have figured out how to create an assembly line without a roof over it. By offering only four or five floor plans in a 200-home development, they create a human conveyor belt:

• The foundation crew hits Lot 1, then Lot 2, then Lot 3.

• The framing crew follows exactly three days behind them.

• The trades (plumbers, electricians) know exactly where every wire goes because they’ve done this exact house 50 times this year.

4. The Rise of “Semi-Smarter” Materials

Since we can’t easily move a whole house, we’ve started moving “smarter parts.” Companies like Builders FirstSource are bridging the gap with products like Ready-Frame.

Instead of a pile of lumber and a saw, the crew gets a pallet of pre-measured, pre-cut, and pre-labeled wood. It’s essentially a “Lego kit” for a house. It reduces on-site wood waste (a major cost sink) and allows a less-experienced crew to assemble a frame with the precision of a master carpenter. This is the “middle ground” of innovation: improving the components because we can’t improve the location.

The Bottom Line

Innovation in homebuilding is slow because we are trying to apply 21st-century logic to a 19th-century constraint: the land is fixed, but the product is heavy. Until we solve the “Logistics of the Bounce” or find a way to standardize building codes across state lines, the only real “assembly line” will remain the one where the workers move and the product stays put. For the small builder, the challenge isn’t just building a better house—it’s finding a way to compete with a machine that they can’t afford to buy.

[I wrote about the demand drivers and supply of the US new home market🏠🏡 in 4/2024, link here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Housing / Banking - , , , ,

• Others-

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

If you like my graphs and charts, this week I am going to direct you to my account on X. I post a lot of my visuals there, and there were too many for me to choose from. I enjoy them all. If you do also, go check them out x.com/NickeNumbers

Podcast- Peter Steinberger | Lex Fridman Podcast

OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman. Peter is the creator of OpenClaw, an open-source AI agent framework that’s the fastest-growing project in GitHub history. He seems like an elite coder to me 🤯. It is a long interview, but facinating.

Video

Queen of Chess👑♟️ is a 2026 Netflix documentary that chronicles the life of Judit Polgár, the greatest female chess player in history. The film, streaming now, highlights her rise from a Hungarian child prodigy to becoming the youngest-ever Grandmaster at 15, breaking barriers in a male-dominated field. And, if you want to be the best, you have to defeat the King🤴🏻, Garry Kasparov. Outstanding 👏🏻

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Best Way to Rob a Bank Is to Own One: How Corporate Executives and Politicians Looted the S&L Industry by William K. Black Business, History.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- Steve Eisman

Steve Eisman is notorious in the financial world, most famous for being one of the few investors who predicted—and profited from—the 2008 subprime mortgage crisis.

Youth and Education- Steven Eisman was born on July 8, 1962 (63 yrs old), into a Jewish family in New York City. He grew up in the city and attended Yeshiva schools, which laid the foundation for his rigorous, questioning analytical style.

• University of Pennsylvania: He graduated magna cum laude in 1984.

• Harvard Law School: He earned his Juris Doctor (JD) with honors in 1987.

Early Career- Despite his legal pedigree, Eisman quickly realized he was “unhappy” practicing law. In a classic Wall Street “foot-in-the-door” move, his parents—who were brokers at Oppenheimer & Co.—arranged a position for him there as an equity analyst. Because of anti-nepotism rules, his parents actually had to pay his first year’s salary themselves. He quickly proved his worth, becoming an “All-Star Analyst” known for his deep dives into specialty finance and the healthcare sector.

The “Big Short” Era- Eisman’s defining moment came during his time as a portfolio manager at FrontPoint Partners (a subsidiary of Morgan Stanley).

• The Bet: Around 2006, Eisman and his team began investigating the “house of cards” that was the subprime mortgage market. They realized that “subprime” loans were being bundled into complex securities (CDOs) that were effectively worthless.

• The Fame: His story was immortalized in Michael Lewis’s book The Big Short. In the 2015 film adaptation, he was the inspiration for the character Mark Baum, portrayed by Steve Carell. The movie captured his “crusader” persona—someone who wasn’t just in it for the money, but was genuinely horrified by the corruption in the banking system.

Career: 2012 to Present- Following the crisis, Eisman’s career had some ups and downs:

• Emrys Partners: He launched his own hedge fund in 2012. However, it struggled to beat the market in a post-crisis world where “fundamentals” didn’t seem to matter as much, and he closed it in 2014.

• Neuberger Berman: He joined Neuberger Berman in 2014 as a Managing Director. He shifted his focus to “old school” industrials, infrastructure, and more recently, the impact of AI.

• Present Day: As of 2025/2026, he has focused his energy on his media presence, launching a weekly podcast titled The Real Eisman Playbook, where he breaks down macro trends and market cycles.

Family and Personal Life- Eisman has been married to Valerie Feigen since 1989. Valerie, a former J.P. Morgan executive, was the basis for the character played by Marisa Tomei in The Big Short. They have two young adult children.

His wife once famously described his bedside manner by saying, “Even on Wall Street, people think he’s rude and obnoxious and aggressive.” Eisman himself seems to agree, once telling an interviewer, “I forget myself sometimes.” ¯\_(ツ)_/¯

Financials and Hobbies

• Estimated Net Worth: Most current estimates place his net worth between $1.5 billion and $1.6 billion. Much of this wealth was solidified by the massive returns generated during the 2008 short.

• Hobbies: Eisman is an avid fan of comic books, particularly the Marvel universe. He has frequently used podcast appearances to critique the narrative quality of modern superhero franchises. Finally, he is a private investor and manages his own money.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about- DRAMpocalypse

The current surge in memory prices (DRAM and NAND) is being driven by a structural shift in the semiconductor industry. As of February 2026, we are in the midst of what analysts call the “DRAMpocalypse,” where AI infrastructure is cannibalizing the production capacity once reserved for consumer electronics.

1. Why DRAM Prices are Rising for Apple & Other Smartphones

For companies like Apple, Samsung, and Xiaomi, the cost of LPDDR5X (smartphone RAM) is projected to rise by as much as 90% in Q1 2026 alone. This is caused by two main factors:

• The “HBM Cannibalization” Effect: High-Bandwidth Memory (HBM) is essential for AI GPUs (like Nvidia’s Blackwell and Rubin). Manufacturing HBM is extremely “wafer-intensive”—it requires roughly 3x the silicon wafer capacity of standard DRAM to produce the same amount of memory. Major players like Samsung and SK Hynix are diverting their limited production lines to HBM because it has much higher profit margins, leaving far fewer wafers for smartphone DRAM.

• The New Competitor (AI Servers): Traditionally, smartphones were the primary users of LPDDR (Low Power) memory. However, Nvidia and other AI chipmakers have begun using LPDDR5X in their AI server CPUs (like the Grace and Vera platforms). This has turned the AI data center industry into a direct competitor with Apple for the exact same pool of memory chips.

2. Why NAND/SSD (Solid State Drive) Prices are Rising for AI Data Centers

While DRAM handles “thinking,” NAND (SSD storage) handles the “storage memory” of AI. The price of NAND is expected to jump 55–60% this quarter due to a shift in how AI models operate:

• Inference & KV Caching: Modern AI models (like GPT-5 or Gemini 3) require massive “KV caches” to maintain context in long conversations. To save costs, data centers push these caches from expensive RAM to high-speed Enterprise SSDs. This has triggered a “gold rush” for high-capacity NVMe drives (64TB and 128TB), wiping out global inventory.

• The Profitability Pivot: Because DRAM is currently so much more profitable than NAND, manufacturers are reallocating some of their NAND production equipment to make DRAM instead. This intentional “choking” of NAND supply, combined with the spike in demand for AI storage, is creating a massive price floor.

Impact Summary: 2026 Market Outlook

The “Long-Term” Reality

While new “megafabs” (like Micron’s New York plant) are under construction, they are not expected to add significant supply until 2028. Consequently, Apple and other manufacturers are forced to either raise retail prices or cut specifications (e.g., sticking with 8GB of RAM when the software really needs 12GB).

🤔- , , , , , , SK Hynix

[Want to take a tour Inside Micron Taiwan’s Semiconductor Factory? I have a a video tour in my post from 4/2025, link here.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , , , , TEL, , , ,

• Housing / Banking -

• Others- , ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast- Elon Musk – "In 36 months, the cheapest place to put AI will be space”

They discuss the economics of orbital data centers, the difficulties of scaling power on Earth, what it would take to manufacture humanoids at high-volume in America, xAI’s business and alignment plans, DOGE, and much more.

…that is on the order of 10,000 Starship launches. Yes. You want to do that in one year. So that’s like one Starship launch every hour. 🤯

Video- Engineering behind stormwater ponds

15 Min. Part park, part plumbing, mostly hiding in plain sight, they are often carefully tuned pieces of infrastructure that help keep the city’s head above water. I am a huge fan of Grady at Practical Engineering because his civil engineering videos are so well done. 👍🏻

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• 1929: Inside the Greatest Crash in History – and How It Shattered a Nation by Andrew Ross Sorkin Business, History, Markets. Fascinating.

• Leadership: In Turbulent Times by Doris Kearns Goodwin History. Abraham Lincoln, Theodore Roosevelt, Franklin D. Roosevelt, and Lyndon B. Johnson. I am trying to make up for not paying attention in high school history class. 😉

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- Do Nothing Investing

When there is nothing to do with your investment choices, you should…. do nothing. I make the large majority of my wealth by… doing nothing.

Buffett talks about a “buy-and-hold” investment philosophy. He believes that the greatest value is often created by doing nothing once a high-quality system is in place.

He says that we should think about investing like an airplane cockpit:

“The cockpit of the future will have only two occupants: a pilot and a dog. The pilot’s job is to sit there and make sure the plane stays on autopilot. The dog’s job is to bite the pilot👀 if he tries to touch any of the controls.”

Once the high quality company stocks are selected the destination is set, and the math of compounding does the work.

Humans have an evolutionary bias toward action, or “fiddling around.” We feel like we aren’t “working” unless we are buying, selling, or adjusting. In the cockpit of a portfolio, this “fiddling” usually leads to errors, taxes, and fees.

The “dog” represents the iron-clad rules/discipline or the temperament required to stay the course. You need a mechanism (like a pre-written investment policy or a partner like Charlie Munger) to stop you from making impulsive “corrections” during market turbulence.

So, when you own high quality companies, and there is nothing to do, do nothing. Or better, go read a book.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about- TSMC 1/14/26 Earnings Call, 5 Key Points ✋🏻

Based on the TSMC 4Q25 earnings call and the subsequent question-and-answer session, my five key takeaways are as follows:

1. Robust Financial Performance and Aggressive 2026 Growth Guidance

TSMC reported a strong fourth quarter in 2025, with revenue reaching $34 billion, slightly exceeding guidance. For the full year 2025, revenue grew by 36% in US dollar terms. Looking ahead, the company provided a highly optimistic outlook for 2026, forecasting full-year revenue to increase by close to 30%, significantly outperforming the broader foundry industry’s projected 14% growth.

2. Significant Increase in Capital Expenditures (CapEx)

To support the “structural demand” from AI and high-performance computing (HPC), TSMC announced a major step-up in capital spending. The 2026 capital budget is set around $54 billion, a substantial increase from the $41 billion spent in 2025.

• Allocation: Approximately 70% to 80% of this budget will be dedicated to advanced process technologies (7nm and below).

• Advanced Packaging: Investment in advanced packaging, testing, and mask-making will account for 10% to 20% of the total budget.

3. Strong Conviction that the “AI Megatrend” is Real

During the Q&A session, Chairman and CEO C.C. Wei addressed concerns regarding a potential “AI bubble”. He stated his conviction that the AI megatrend is “real” and potentially “endless” for many years to come.

• Customer Evidence: Wei confirmed that cloud service providers have shown evidence of real financial returns and business growth driven by AI.

• Forecast Raise: TSMC raised its CAGR forecast for AI accelerator revenue growth to mid-to-high 50 percent for the five-year period from 2024 to 2029.

4. Successful Launch of 2nm (N2) and Future Node Roadmap

TSMC’s latest leading-edge technology, N2 (2-nanometer), successfully entered high-volume manufacturing in 4Q 2025 with good yield.

• Rapid Ramp: The company expects a faster ramp-up for N2 in 2026 due to strong demand from smartphone and HPC/AI customers.

• Future Nodes: Volume production for N2P (an enhancement of N2) and A16 (featuring “superpower rail” technology) is scheduled for the second half of 2026^.

5. Acceleration of Global Manufacturing Expansion

TSMC is speeding up its overseas expansion, particularly in Arizona, to meet intense customer demand^.

• Arizona Pull-forward: Production for the second Arizona fab has been pulled forward to the second half of 2027.

• GIGAFAB Cluster: The company has purchased additional land in Arizona and is applying for permits for a fourth fab and its first advanced packaging fab in the U.S., aiming to create an independent “GIGAFAB cluster”.

• Global Progress: The first specialty fab in Kumamoto, Japan, began volume production in late 2024 with “very good yield,” and construction of a second fab there has already started.

Think🤔- , , , , , , ,

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Housing / Banking - , , , ,

• Others-

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Video

30 Min, Satya Nadella on AI’s Business Revolution: What Happens to SaaS, OpenAI, and Microsoft?

[If you want to learn more about Microsoft, I wrote about its CFO, Amy Hood, in my 10/2023 post. She is one of my favorite CFOs. Link here.]

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Water Knife by Paolo Bacigalupi Environmental Fiction, Apocalyptic. This was a recommendation of . Outstanding, 5 ⭐.

• Typhoon by Joseph Conrad Classic Fiction. Recommendation by CEO Ted Sarandos. Yep, it was very good.

• Use of Weapons by Iain M. Banks Science Fiction, Classic. Fantastic.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏-Stock Market Circuit Breakers

In the last 10 years (2016–2026), the U.S. stock market has experienced a range of volatility that put its circuit breaker systems to the test. While single-stock halts (Limit Up-Limit Down) occur almost daily, Market-Wide Circuit Breakers (MWCB)—which halt the entire New York Stock Exchange (NYSE) and Nasdaq—remain exceptionally rare.

The Three-Level Threshold System

Current rules (established in 2013) trigger halts based on the S&P 500 Index’s decline from the previous day’s close:

• Level 1 (7% drop): 15-minute trading halt (if before 3:25 PM ET).

• Level 2 (13% drop): 15-minute trading halt (if before 3:25 PM ET).

• Level 3 (20% drop): Trading halted for the remainder of the day.

CHRONOLOGICAL HISTORY of MARKET DROPS (2016–2026)

2016–2019: The “Safety Net” Era

Despite periods of volatility—including the 2016 Brexit vote and the 2018 “Volmageddon” spike—no market-wide circuit breakers were triggered. During this time, the system was viewed as a background safety mechanism that many believed would never be used in a modern, high-speed trading environment.

March 2020: The “COVID Crash” Cluster

The most significant period in the history of circuit breakers occurred in March 2020. The system was triggered four times in ten days, the first instances since 1997 (22 years prior.)

• March 9, 2020: Triggered minutes after the open due to oil price wars and escalating pandemic fears.

• March 12, 2020: Triggered again as the NBA suspended its season and travel bans were announced.

• March 16, 2020: Triggered immediately at the opening bell after the S&P 500 fell 7% in seconds.

• March 18, 2020: The final Level 1 trigger of the era occurred in the afternoon.

2021–2023: Bear Market Resilience

The 2022 bear market saw the S&P 500 drop nearly 20% over the year, but the decline was gradual. No single day saw a 7% drop, proving that circuit breakers are designed for “flash” panic rather than long-term economic downturns.

2024–2025: Technical Glitches and Near Misses

• June 3, 2024: A technical glitch at the NYSE caused hundreds of individual stocks (including Berkshire Hathaway) to show 99% drops. While this triggered thousands of Limit Up-Limit Down (LULD) halts, it did not trigger a market-wide circuit breaker.

• August 5, 2024: Following a “carry trade” collapse in Japan, the Nikkei 225 hit its own circuit breaker (12.4% drop). The U.S. S&P 500 dropped roughly 3%—significant, but well short of the 7% threshold.

• April 4, 2025: Amid “Tariff Turmoil” and trade war concerns, the S&P 500 plummeted 6% in a single session. This was the closest “near miss” for a Level 1 trigger since 2020. International markets in Taiwan and Japan did hit their triggers during this week.

Early 2026: Current Status

As of January 2026, the U.S. markets have entered a recovery phase with lower volatility. No market-wide halts have been recorded since the 2020 cluster.

SUMMARY OF MAJOR MWCB TRIGGERS

Date Trigger Level Primary Cause

March 9, 2020 Level 1 (7%) COVID-19 pandemic / Oil price collapse

March 12, 2020 Level 1 (7%) Rapid escalation of pandemic lockdowns

March 16, 2020 Level 1 (7%) Extreme liquidity panic at the opening bell

March 18, 2020 Level 1 (7%) Continued uncertainty and margin calls

April 4, 2025 Near Miss (6%) Global trade war/tariff policy announcements

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Bureaucracy in Society 👎🏻

The New York Times had an article in October 2025- “Opening a Restaurant in New York Is No Picnic. Here’s What It Takes” (by Priya Krishna). It is a sobering look at the extreme bureaucratic, financial, and logistical hurdles required to launch a new restaurant in New York City.😬😫

Here is a summary of some of the key challenges and timelines:

1. The “Dead Rent” Trap- The biggest financial killer for new restaurateurs is paying rent while the doors are still closed. Because of the lengthy permit process, owners often pay months—or even over a year—of “dead rent” (sometimes $20,000–$50,000+ per month) before serving a single meal. Most landlords offer a “rent abatement” period, but it rarely lasts long enough to cover the actual time it takes to open.

2. The Bureaucratic “Gauntlet”- Opening requires approval from a dizzying array of city agencies that do not coordinate with one another. To open, you must satisfy:

• The Department of Buildings (DOB): For construction permits and Certificates of Occupancy.

• The Department of Health (DOH): For food safety and kitchen design.

• The FDNY: For fire suppression systems (which can take months to inspect).

• The Landmarks Preservation Commission: If the building is in a historic district.

3. The Con Edison Struggle- One of the most cited “nightmares” in the article is getting gas service turned on. After the 2015 East Village gas explosion, regulations became significantly stricter. It is now common for restaurants to wait six months to a year for Con Edison to inspect pipes and turn on the gas, forcing many to pivot to all-electric kitchens at the last minute or cook on induction burners.

4. The Liquor License Battle- Securing a liquor license is a two-step ordeal that can take a year:

• Community Boards: Owners must first present to local community boards, which can be hostile and demand early closing hours or “no-music” stipulations.

• The State Liquor Authority (SLA): Even with board approval, the SLA has a massive backlog. Without a liquor license, the profit margins for most NYC restaurants are non-existent.

5. Hidden Costs and Professionals- The article emphasizes that you cannot do this alone. To navigate the red tape, owners must hire:

• Expediters: Fixers who specifically know how to navigate the DOB and SLA paperwork.

• Specialized Lawyers: To handle the lease and liquor board.

• Architects and Engineers: Who understand specific NYC fire and health codes.

The startup budget for a small, 30-seat restaurant can easily balloon past $500,000 before a single customer walks in.

6. The “Human Toll”- Beyond the money, think of the psychological exhaustion. Many owners exhaust their life savings, take out predatory loans, or experience mental health crises due to the uncertainty of when (or if) they will be allowed to open.

The Bottom Line

While New York remains a world-class food capital, the barrier to entry has become so high that independent, “mom-and-pop” restaurants are being squeezed out, leaving the market primarily to large hospitality groups with the deep pockets necessary to survive the 12-to-18-month “pre-opening” phase.

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- at CES

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Look at the intra-year declines📉. Most years have them. Opportunity 😀!

Spot GPU prices at AWS rising across all GPU vintages for the last few months. 🤔

Podcast- IBM, Mixture of Experts

Excellent episode with outstanding opinion/perspective. ⭐👍🏻

CES 2026 , Meta’s $2 Billion Manus bet , new DeepSeek paper tackles scaling , AI optimism vs. fear

Video- 60 Minute Archive- Howard Buffett

13 Minute video, profile of farmer and American philanthropist Howard Buffett, the son of Warren Buffett. TBH, I was impressed with his intelligence and experience.

Howard Buffett: “It seemed nothing I could do would be as successful as what he did, meaning Dad... That’s okay. I mean, you know, my mom and dad always made it clear that that was okay.”

Warren Buffett: “Having Howie there adds just one extra layer of protection... he knows the values of it [the business].”

[FWIW, I think Berkshire Hathaway has WAY (WAYYYyyy) too much Cash. And I wrote about what to do with it in my 8/2024 post, here.]

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Who Knew by Barry Diller Biography, Business

• The Unaccountability Machine: Why Big Systems Make Terrible Decisions - and How The World Lost its Mind by Dan Davies Business, Economics

• Titanium Noir by Nick Harkaway Science Fiction. Different and original. ⭐

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏 The “Hot Water” Paradox: Jensen Huang’s CES 2026

At the recent CES 2026 keynote (link here), NVIDIA CEO Jensen Huang made a statement that sounded like a typo to many: the water flowing into their newest data center racks is now 45° C (120° F).

To a human, that’s the temperature of a hot shower. To a traditional data center, it’s a radical departure from the “refrigerator” mindset of the last two decades. For years, we’ve been taught that computers need to be kept in freezing, air-conditioned rooms. Jensen is signaling the end of that era, moving toward what he calls the “Physical AI Era” where liquid cooling is no longer optional—it is the foundation.

The Physics of the Thermal Gradient- How can you cool a high-performance GPU with water that is already hot? The answer lies in the Thermal Gradient.

A modern AI chip, like the Blackwell (GB200) or the newly announced Vera Rubin systems, generates an incredible amount of heat. Under full load, these chips can reach internal temperatures of 80-95°C (180°F.)

• Because heat naturally moves from “hotter” to “colder,” the 45°C water acts as a massive thermal sponge.

• Since water is roughly 4,000 times more effective at carrying heat than air, it can pull 40°C worth of heat away from the chip almost instantly.

• This process happens through Direct-to-Chip (DTC) cold plates—micro-channeled copper blocks that sit directly on the processors, allowing the water to whisk heat away without the need for noisy, inefficient fans.

Killing the Chiller: The “Before” vs. “After”- The true brilliance of 45° C water isn’t just about the chip; it’s about the building.

Previously, data centers used chilled water systems where the inlet temperature was typically between 7-15° C. To get water that cold, you need massive, power-hungry mechanical chillers—essentially giant industrial refrigerators. These chillers often accounted for 30% to 40% of a data center’s total electricity bill.

By raising the water temperature to 45°C, NVIDIA has “killed the chiller.”

• Even in a hot climate where the outside air is 35°C (95°F), that air is still “cooler” than the 45°C water.

• This allows data centers to use Dry Coolers (massive radiators with fans) to reject heat into the atmosphere using nothing but ambient air. This shift is so significant that stocks for traditional chiller manufacturers saw a sharp decline following Jensen’s speech.

Turning Waste into Wealth: Heat Reuse- One of the most exciting aspects of this shift is what happens to the water after it leaves the rack. When the water enters at 45°C, it might exit at 60°C or higher.

This “waste” water is no longer just a nuisance to be cooled down; it is a high-grade energy source.

• District Heating: Data centers are now being integrated into city grids to provide hot water and heating for homes and offices.

• Energy Efficiency: This pushes the Power Usage Effectiveness (PUE) from a legacy average of 1.6 down to 1.1 or lower. In this new model, 90% of the electricity goes to actual AI computation, while only 10% is “lost” to the facility operations.

Conclusion: Scaling to the Million-GPU Frontier

Why does this matter now? Because we are reaching the limits of air. As AI models move toward trillions of parameters, we are building “AI Factories” with hundreds of thousands of GPUs in a single cluster.

The move to 45°C water cooling removes the “thermal ceiling” that was threatening to stall AI progress. We are moving away from building data centers that fight against the environment and toward systems that work with the laws of thermodynamics to scale more sustainably.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Invention/Discovery vs. Reverse Engineering Technology

The difference between original invention and reverse engineering can be better understood by thinking about the search space and the existence proof. When you are inventing something from scratch, you aren’t just looking for a needle in a haystack—you don’t even know if the haystack contains a needle, or if needles can even exist in hay.

Here is why that psychological and technical gap is so massive:

1. The “Existence Proof” (Psychological Certainty)

The moment you know someone else has done it, the nature of the problem changes from “Is this possible?” to “How did they do it?” Think about Chinese competitors attempting to copy ’s EUV process, for instance. The physical reality hasn’t changed; the certainty of the goal has.

• Cognitive Fuel: In original invention, “failure” feels like a dead end (maybe the idea is just impossible). In reverse engineering, “failure” is just a temporary lack of understanding. You persevere longer because you know a solution exists.

2. Unbounded vs. Bounded Search Space

Invention requires navigating an unbounded search space. If you are trying to create a large language model, your “hypothesis” could involve anything from videos of humans talking and reading books to directing a transformer algo to crawl the entire internet.

• Original Invention: Create a “self-attention” mechanism that allows neural networks to analyze entire bodies of text simultaneously. You must decide which variables even matter.

• Reverse Engineering: The search space is bounded by the results in front of you. You can see the input window, the dimensions, and the outputs. Even if the internal mechanism is a “black box,” you can measure what goes in and what comes out. You are no longer guessing what to build, but how the existing pieces interact.

3. The Burden of “False Positives”

When inventing, many hypotheses look promising but lead to “local maxima”—solutions that work a little bit but can never be scaled or perfected. You might spend years on a dead-end path because you don’t know there’s a better way.

In reverse engineering, you are looking at a global maximum. You know you are aiming for a version that is already optimized for the real world, which prevents you from getting stuck in “good enough” traps. (LLM Leaderboards and test results.)

[A suggestion for 2026 on independent thinking and how I consume financial statements in a throwback post from 3/2024, link.]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- ,

• Housing / Banking - ,

• Others- ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Video- OpenAI is King of Losing Money? Warning From Jim Chanos

Chanos breaks down why hosting GPUs is a commodity business with low returns and why the depreciation of AI chips (like Nvidia’s) creates a massive financial risk for companies like CoreWeave and Oracle. He also discusses the dangers of private credit, the accounting tricks at Live Nation, and why the “unprofitable” nature of today’s AI customers makes this cycle riskier than the Dotcom era.

“Private Credit Reminds Me of Michael Milken”

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Breakneck: China’s Quest to Engineer the Future by Dan Wang Nonfiction, Economics, China. Interesting read.

• The Storied Life of A.J. Fikry by Gabrielle Zevin Fiction. Outstanding writing.

• The Universe Speaks in Numbers: How Modern Maths Reveals Nature’s Deepest Secrets by Graham Farmelo Nonfiction, Mathematics, History

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏

Large national homebuilders (like D.R. Horton, Lennar, and Pulte) prefer "velocity" (speed and volume) over "margin" (profit per house). Why?

Large national homebuilders have fundamentally shifted their business models over the last decade. They no longer view themselves as real estate speculators who wait for land to appreciate; instead, they view themselves as Industrial Manufacturers. , ,

1. Asset Turnover and ROI (The Grocery Store Model)

In finance, there are two ways to make a lot of money: have a high profit margin on a few items (like Ferrari) or a tiny profit margin on millions of items (like Walmart).

• The Math: If a builder makes a 20% margin on 100 homes that take two years to sell, their Return on Investment (ROI) is slower than a builder who makes a 10% margin on 400 homes in the same two years.

• The Result: By moving inventory faster, they can reinvest that capital into new land and new projects. Large builders prioritize Internal Rate of Return (IRR) over simple profit margins.

2. High Holding Costs (”Carry”)

Holding onto a lot (land) is expensive. Every day a house isn’t built, the builder is losing money to:

• Property Taxes: They pay taxes on the land regardless of whether a house is on it.

• Financing Costs: Most builders use massive lines of credit. They pay interest on the money used to buy and develop the land.

• Opportunity Cost: Money tied up in a “reserved” lot is money that isn’t being used to buy the next piece of land in a hotter market.

3. Maintaining the “Machine” (Labor and Supply Chain)

Homebuilding requires a massive, complex network of subcontractors (plumbers, framers, electricians).

• Labor Loyalty: Subcontractors want steady, predictable work. If a national builder slows down to “wait for higher prices,” those crews will go work for a competitor. Once a builder loses their reliable labor force, it is incredibly difficult and expensive to get them back.

• Volume Discounts: Builders get massive discounts from suppliers (like Whirlpool for appliances or Mohawk for flooring) because they promise to buy thousands of units per quarter. If they slow their pace, they lose their “bulk-buy” bargaining power, which actually raises their costs.

4. Stock Market Pressure

Most large builders are publicly traded companies. Wall Street punishes “lumpy” earnings.

• Predictability: Shareholders want to see consistent quarterly growth in closings and revenue.

• Absorption Rates: Analysts track “absorption”—how many homes a builder sells per community per month. If that number drops because a builder is “holding out,” the stock price usually tanks, even if the eventual profit per home is higher.

5. The “Land Light” Strategy

In the past, builders bought huge tracts of land and held them for years. Today, many use Option Contracts.

• They pay a small fee to a land developer to “reserve” lots, but they don’t actually buy the lot until they are ready to start the foundation.

• This means they don’t “own” the land in the traditional sense; they just have the right to build on it. This forces a “use it or lose it” mentality—if they don’t build at a certain pace, they lose their options on the future lots.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

Income Tax Differences of Trading vs. Investing-

The end of the calendar year also marks the end of another tax year 💸🫰🏽. I think of myself as 95% long term investor and 5% trader, but the tax consequences of the two styles are dramatically different.

Warren Buffett made the following statement on income tax differences between Trading and Investing in his 1989 Letter to Shareholders of Berkshire Hathaway-

Imagine that Berkshire had only $1, which we put in a security that doubled by yearend and was then sold. Imagine further that we used the after-tax proceeds to repeat this process in each of the next 19 years, scoring a double each time. At the end of the 20 years, the 34% capital gains tax that we would have paid on the profits from each sale would have delivered about $13,000 to the government and we would be left with about $25,250. Not bad.

If, however, we made a single fantastic investment that itself doubled 20 times during the 20 years, our dollar would grow to $1,048,576. Were we then to cash out, we would pay a 34% tax of roughly $356,500 and be left with about $692,000.

The sole reason for this staggering difference in results would be the timing of tax payments. Interestingly, the government would gain from Scenario 2 in exactly the same 27:1 ratio as we, taking in taxes of $356,500 versus $13,000. Though, admittedly, it would have to wait (20 years) for its money.

The results seem so drastically different, I find myself questioning Uncle Warren. So, in order to confirm his math, I dropped the 2 scenarios into a spreadsheet and… Voila! Warren is 100% correct.

Trading with a tax on the left. 👈🏽

Long term Investing with a deferred tax on the right.👉🏽

[Income Taxes can be brutal. Consumer Retail is also brutal and I wrote about it in 8/2023, here. Think Target, Kohls, Best Buys. Someone often goes bankrupt after Christmas🎄. Beware⚠️]

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- ,

• Housing / Banking -

• Others-

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Google owns 7% of SpaceX. Lets goooo!!!

So far in 2025, and with an entire quarter to go until the end of the year, Cloudflare has already mitigated 36.2 million DDoS attacks. That corresponds to 170% of the DDoS attacks Cloudflare mitigated throughout 2024.

About 7% of New Homes Are Teardowns (structures torn down and rebuilt in older neighborhoods).

If Americans are reading the news less and passively investing more, I suspect successful active investors can achieve more alpha. 😃🥳 Sad but true.

Podcast- This Is Why Credit Card Interest Rates Are So High💳

Some people pay off their credit cards at the end of each month. They use the cards as a payment method and collect points and rewards, and never have to pay any interest. For other users (revolvers), interest can be sky high — way higher than what would be expected simply based on a user’s credit or default risk. So why is this? And how do credit card companies get away with charging interest at these levels? Itamar Dreschsler, a finance professor at Wharton, walks us through the costs of running a credit card operation, and explains what borrowers are really paying for.

“...the big part of their business where most of the money comes from... is all these people that revolve, they pay an interest rate and that interest rate now is on average 23%. Which is a shockingly high number...”😮

You could say well why doesn’t somebody just cut all this marketing out and just charge a lower rate and that’ll get people?... Apparently it doesn’t work. People are not rate sensitive...”🤦🏽🤦🏻‍♀️

Video- Gavin Baker is ⭐

Patrick O’Shaughnessy sits down with investor Gavin Baker to explore the rapidly evolving AI landscape. They dive deep into the infrastructure war between Nvidia and Google, discuss the implications of Gemini 3 and scaling laws, and examine how the transition from Hopper to Blackwell chips is reshaping the industry. Baker shares his insights on frontier AI models and the economics of token production. The conversation also covers data centers in space, the future of robotics, and why traditional SaaS companies are making critical mistakes with AI adoption. A comprehensive look at the technical, economic, and strategic forces driving the AI revolution. 🔥 , ,

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• The Essays of Warren Buffett: Lessons for Investors and Managers by Lawrence A. Cunningham Business

• Lost & Found: A Memoir by Kathryn Schulz Nonfiction, Memoir.

• A Thousand Brains: A New Theory of Intelligence by Jeff Hawkins Science, Nonfiction.

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏 Michael Burry Permabear

Michael Burry, one of the investors immortalized for correctly predicting the 2008 housing crisis in The Big Short, has gained a reputation that makes his latest warnings about an AI bubble feel both profound and profoundly repetitive. I don’t know Burry myself, and he is probably a fine human. But, as an investor he has become a figure akin to the boy who cried wolf too many times, or the imagery that I prefer is the man who calls fire in a crowded theater, often.

While his initial, meticulously researched call on the subprime mortgage crisis was a rare moment of clarity that earned him a fortune and legendary status, his subsequent track record is one of constant, often premature, and unfulfilled market pessimism. I have kept an eye on him over the last 4 years and he cries fire about 2 times per month. That is about 96 Michael Burry fire alerts over the last 4 years. At this point I think he is crazy. To be specific I think Burry is crazier than a cat🐈‍⬛ trying to bury a turd on a hardwood floor.

Since the Global Financial Crisis, Burry has issued numerous dire predictions—ranging from a 2017 global financial meltdown to warnings about index fund bubbles, and repeated calls for stock market crashes in 2019, 2021, and 2022. For an investor constantly predicting “doom,” the market has often continued to climb, penalizing those who might have followed his advice to sell. This consistent pattern of pessimism, regardless of the immediate market’s direction, has led critics to label him a “permabear”—someone who always expects the worst.

People who are new to investing, watch The Big Short once, and hear his latest predictions think they have found some investing edge. Anyone who has listened to Burry for six months would know that his statements are like a goat arguing with a car alarm. 🚨+🐐 (Distracting noise.)

He cried fire last week with a position on the AI sector, where he is actively betting against heavyweights like Nvidia and Palantir. His opinion is undoubtedly based on detailed analysis, pointing to concerning factors like stretched valuations, an “oversupply crisis,” and what he views as questionable accounting practices (specifically the extended depreciation of expensive AI hardware). However, for anyone that has watched fruitcake Burry for any length of time, the message of an impending bubble is no longer a novel insight; it’s just the latest installment in Burry’s consistent Nutty narrative.

The fame derived from The Big Short ensures that his warnings instantly generate headlines and fear, granting his latest pronouncements an undue sense of urgency. Ironically, this media amplification of his every pessimistic view is more a motivation for media to get buzzy engagement📰🗞️📰, then to actually hold Burry accountable to This prediction or any of his many previous ones.

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment

Subscribe now

Something that I am thinking about

The US Cement Industry

I was the CFO for a heavy industrial infrastructure manufacturer. Think quarries, asphalt, concrete along with civil engineering (roads, and bridges.) I have a passion for technology and semiconductors, but also for heavy infrastructure. That being said, I am thinking about the US Cement Industry.

First of all, cement is not🙅🏽‍♂️ concrete. Cement is the powdered binder that acts as the “glue” when mixed with water, while Concrete is the final composite material made by mixing cement with aggregates (like sand and gravel) and water. Think of cement as a single ingredient in a recipe, and concrete as the finished dish that contains many ingredients.

The U.S. domestic cement supply chain consists of ~99 operational plants situated across 34 states and Puerto Rico. Production remains heavily concentrated, with Texas, Missouri, California, and Florida together accounting for approximately 43% of total U.S. output. (Only 99 plants in the whole USA🤔, seems like it would be more.) Since permitting and building new conventional plants takes a decade or more, the industry’s immediate supply expansion capability is severely limited. Largest providers- (formerly Holcim), Heidelberg, , , , .

Supply of Cement-

Pricing trends over time.

Consumption by Major End-Use Sector-

Long-Term Trends-

Long-term projections highlight a severe structural capacity issue: the U.S. faces a materials supply gap of up to 50 MMT of cement by 2050—approximately 43% increase from current production levels (think new electric generation, new electric transmission and new data centers builds.) Imported cement faces high tariffs(~10-25%) and high transportation costs. This deficit is magnified by the difficulty and time commitment required to permit and build new conventional cement plants.

Increasing existing cement plant production capacity, incremental efficiency and materials science is necessary to bridge the anticipated supply deficit and meet the nation’s massive infrastructure and housing needs.

I follow the public cement producers listed above and hold positions in some of them. I also follow the public quarry companies, like , , and .

Investing, Companies, Market Past/Future

(I invest in Companies, not in stocks.)

Interesting companies that hosted earnings results or information sessions since my last edition that I reviewed-

• Technology- , ,

• Others- ,

(With data and charts it is not what you look at, it is what you see that matters.) Click on any of the images below to see a larger view.

Podcast- Why It’s So Difficult For Companies to Go Public: A CFO Explains the Process

On this episode of The Real Eisman Playbook, Steve Eisman is joined by Amy Butte, the Chief Financial Officer at Navan. Amy was also the CFO of the New York Stock Exchange when it went public in the early 2000s. They explore what it means and how difficult it is to take a private company public.

Video- Ilya Sutskever – We’re moving from the age of scaling to the age of research

Ilya is the OG of ML, LLMs, Gen AI, etc. His resume is platinum🪙🪙. The discussion between Dwarkesh and Ilya is fascinating. One of the things that I like most about Ilya is that he thinks and talks in analogies and this makes his subject and his thoughts much easier to grab a hold of. I can’t wait to see what Ilya does next with his research and technology. Enjoy!

Up until 2020... it was the Age of Research now from 2020 to 2025 it was the Age of Scaling... but now the scale is so big... it’s back to the Age of Research again just with big computers.”🤯

Books and Articles since my last edition

(Reading Great books is like a conversation with the finest minds of the past.- Descartes. If I get bored of a book I quit and move on.)

• Who Is Government? The Untold Story of Public Service by Michael Lewis Nonfiction

• The Cuckoo’s Egg: Tracking a Spy Through the Maze of Computer Espionage by Clifford Stoll Nonfiction, Cyber Crime

Here is a link to many of the books I have read and my 1-5 star rating- Goodreads Books Read

Wild Card🃏- More is Not more. Less is more.

I am writing about less is more in the context of investing/finance but what is good for me might not be good for you. As my years of experience have rolled by I don’t think more positions, more opinions or endless investing input is valuable. I mean my substack name and opening sentence from day one has been “In a busy world we can be distracted with the trivial many things or we can focus on The Vital Few.” But too often I see investors substitute activity for productivity and constant repetitive chatter instead of quiet reflection.

This reminds me of a Nobel Prize winner Richard Feynman’s comment-

A large part of wisdom is knowing what to ignore. A large part of expertise is knowing where to place your attention.

How much investing news can we listen to? Endless. How many public companies can we attempt to invest in? 500 in the S&P and more than 4000 US public companies. When have we performed enough research and analysis? More, and more. I know investors that spend time and significant money flying to visit companies, competitors, countries and investing conferences. Is there significant ROI on this time/travel cost spend? It feels good to be busy😵‍💫. Kids like candy 🍬🍭.

My challenge to myself and suggestion for you is to look for things to remove from your process. Low value news, duplicate opinion providers, delete or unsubscribe. Things that distract or cause you to spend your work time unproductively, delete.

If you are a math person, you can think of it as Addition by Subtraction. My investing signal📡 or focus gets clearer the more nonsense noise I remove.

When Michelangelo carved David from a block of stone he didn’t go around and ask 27 people what he should carve. He didn’t line up 14 stone blocks and knock a few bits off each of them never progressing. He focused on one block of stone for only 3 years, each day removing a little more. Michelangelo said something like:

Every block of stone has a statue inside it and it is the task of the sculptor to discover it.

2025 has been a great year for me in investing and I own less than 15 positions. Depending on what your investing goals are and what your experience level is don’t confuse yourself into spending lots of time and activity on everything and thus nothing. Find the things to begin subtracting from your process and focus more on the vital few important things that provide the most signal.

[If you liked this section, check out my 10/2023 post here, where I talk about primary vs secondary information sources.🤔 Cheers!]

If you liked this post from Nick, why not share it and subscribe?

Share Nick’s Vital Few

Subscribe now

Disclaimer: All of my posts are for informational purposes only. I might own some of the companies discussed in these posts. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.

Leave a comment