I’m Francesco, and I invest in the backbone of the future with Visionaries Tomorrow. If you are crazy enough to go full-stack in robotics, hit me up!

Subscribe now

Some years ago a few friends and I decided to try our luck and embark on the crazy journey of entrepreneurship. We were in the midst of the quick-delivery bubble and billions were being invested in anything related to it. Back then we were good roboticists and nothing more than that: we only had one hammer (we could build robots) and so every problem looked like a nail to us, including that of transforming a money losing industry into a profitable one.

All of a sudden, we were in the business of building mobile robots to deliver food.

After putting our heads on the problem, we soon decided it made no sense to build our own hardware when we could buy an off-the-shelf Chinese product for much less, and have it delivered to us industrialized and working. We believed that robotics hardware was becoming a commodity, and there was no value in building a better chassis: we thought that the most important layer was the software steering the mobile base, and that we wanted to build that.

Searching on the web, we realized that a Chinese firm had bought the once-glorious brand Segway and had transformed it into a robotics company - Segway Robotics. Soon enough we had a quote in our inbox, and they were ready to ship the product the day after.

However, we wanted to be the European Coco Delivery (they had just raised $36M and we wanted to be as cool as them), so we didn’t go with Segway and bought the same chassis Coco was using (a more expensive, US-made, Rover Robotics).

Fast forward to today, we didn’t manage to change the world of food delivery - but something more interesting happened.

Segway moved up the value chain: today they are not selling bare chassis anymore, but full mobile robots, with integrated sensors, screens, and - surprise, surprise - a proprietary software stack.

https://robotics.segway.com/e1/

And even more interesting, guess who is listed as a customer?

The good old Coco Robotics.

The one used by Segway is the same playbook Chinese technology manufacturers have used to take over many of the industries that once defined the West.

It all starts with a low value-added, seemingly low IP, step of the value chain. Assembling is often a good point to start: for Segway Robotics it was assembling cheap electric motors on a plastic chassis.

Slowly, but steadily, these companies start moving up the chain, capturing increasingly higher value-added layers. Each one of these steps looks like a no-brainer to the end customer: why build your own chassis when theirs is cheaper and ready to buy? why spend months on sensor fusion algorithms if you can have them off-the-shelf and pre-calibrated?

The problem is that each one of these decisions is correct, if taken in isolation. In the short term, at a company level, they actually are no brainers. It’s only when thinking long-term, and at industry or continental level, that we realize this is exactly how industries slowly change hands.

Why does this keep happening?

If China is involved, many will point to cheap labor or State incentives. While this can play a role, I believe the reason is more structural.

When a new industry is born, everyone is vertically integrated because there is no alternative: there is no supply chain to buy from.

As long as the industry is nascent, external suppliers won’t invest in it. One obvious reason is that no one wants to invest to supply an industry that might not exist next year. But the most important reason is that the interfaces between stages aren’t defined yet, so you can’t hand a clean spec to a vendor for something you’re still inventing. Not having defined interfaces also means that uncertainty over the final product would flow through the whole chain and so when the product changes every quarter, suppliers just can’t keep up.

This is the beauty and the curse of vertical integration: it’s hell to deal with, but your product innovation is not slowed down by any supplier. Your fate is entirely in your hands.

When the industry matures, the logic inverts. As the product settles and the uncertainty drains away, specialization plus scale beat vertical integration. With a mature supply chain, each specialist gets its own economies of scale and competition grinds each margin down.

Long and fragmented supply chains give us cheaper products, but make industries react slowly, because anything new has to ripple through the whole chain.

While this is not a problem in mature industries at steady state, it can be lethal during times of change. When new technologies arrive, or adjacent industries start invading the market, all of a sudden the mature industry is thrown back at station zero, and all interfaces are up for grabs again. This is where vertical integration wins.

This cycle vertical integration → fragmented supply chains → vertical integration is known in academic circles as “double helix”. It is exactly when things are not static anymore and the helix starts turning that new entrants can make their way in and take over old incumbent industries.

A long supply chain has inertia. No one owns the whole product, so no one can redesign it end to end, and every interface is a wedge.

New entrants hardly storm the front gate, but instead come in at the edges of big markets (USV docet!). They start from the unglamorous corners, the low-value-added tasks incumbents give away with pleasure, learn the chain one link at a time and integrate as they go.

This is how China moved from being marginal in an Internal Combustion Engine-dominated automotive supply chain, to being dominant in an EV-dominated one. China is today the world’s top exporter of cars (up 35x in the last 13 years) and the leading firms like BYD are heavily vertically integrated - around 75% of the parts made in-house.

This is the story of Segway Robotics. This is also the story of the broader robotics space: an industry that was once dominated by German, Japanese and Italian firms when “robot” meant “big industrial arm”, and that is undergoing technological shifts that make the new interfaces up for grabs.

Chinese firms, as you would expect, are using this wedge to take over the industry: ~87% of the world’s humanoids shipped in 2025 were Chinese. And speaking of vertical integration, you shouldn’t be surprised by now to learn that Unitree, the world’s leading humanoid firm, produces 86% of its components in-house.

What would’ve happened to our little robot delivery company if we had not given up?

We would have built a beautiful navigation stack on somebody else’s chassis. Then they’d have added sensors, and after a while we’d have deleted our sensor work. Then a screen and a fleet dashboard, and a navigation stack, and on and on.

At which point we would have been a system integrator with a European sales team.

At the end of the day, isn’t it the same trajectory European car makers are on? And isn’t it the same trajectory most European robotics startups are on as well?

As I have clearly said, I feel you - when facing the decision, choosing to buy instead of building does look like a no-brainer. Unitrees are amazing products, and vertically integrating looks scary as hell.

But ask yourself the question: what do you want to be, in the long run? Do you want to be a system integrator, importing Chinese robots and dealing with your customers’ WiFi firewalls and MES integrations at most?

If history repeats, and it does repeat, this is exactly what is going to happen.

Obviously this is not bad by definition, but I don’t believe it is the path to building a generational company.

The alternative path, vertically integrating and fighting an open battle, is much harder - but at least it is a battle worth fighting because the prize for winning is much bigger.

And the thing is, you don’t have to vertically integrate forever. As we have seen before, no market stays vertically integrated forever. But by skipping the vertical integration step altogether, we fail to create the conditions for a long supply chain to exist - which leaves us dependent on a kind of provider that knows exactly what its desired role in the world is, and knows the playbook to achieve that.

This is not the future we want, and not the future we are investing in.

Subscribe now

I’m Francesco, and I invest in the backbone of the future with Visionaries Tomorrow. If you are crazy enough to go full-stack in robotics, hit me up!

Why is that? And where is the Revolut of Energy?

Before I opened my Revolut account, I had never really considered banking as something you can do from your phone. Sure, I could check my current balance, and maybe do a clumsy wire transfer, but banking belonged to a physical location, or to the laptop at best. There was certainly no easy way to share money at the tap of the finger, or to immediately check whether the supermarket purchase went through.

While banking apps existed, they were built on top of an infrastructure that belonged to the previous century. Revolut, for the first time, was a product built for an age in which the smartphone was always in customers’ hands.

This seemingly little detail generated what’s today a €100b+ juggernaut. What mobile really did is to change what counted as a banking event. Splitting the bill or exchanging currencies were not banking events before, because the friction to do those things through the bank was prohibitive. Mobile collapsed the friction, and the volume of banking events per customer went up orders of magnitude. With that volume came data, and with data came the ability to insert adjacent products. With that, came the position Revolut now occupies: the substrate of every financial decision its customers make.

It is hard to overstate how much of that story is downstream of mobile. Without the smartphone in 2007, none of the neobanks would have happened.

I think energy is in the same position now, and that the home battery is doing for energy what the smartphone did for banking.

My relationship with utilities today is similar to that with my bank pre-Revolut. A bill arrives at the end of the month, summarizing a consumption that already happened and without the ability to act on any data point - equivalent to the monthly bank statement in financial services. Energy is a real-time market, that sends signals no customer can hear, because no customer is in the position to act on them.

Conditions are now mature for this to change.

Electricity is the cheapest form of final energy, and battery prices have fallen over 80% in a decade and continue to fall. But while the cost of producing an electron went down, the cost of delivering it has risen: over the past decade in the US, generation costs fell 33% while delivery costs rose 65%, and the same dynamic is playing out across Europe at different speeds. Network charges and levies now make up the majority of the household bill in Germany. The grid as it is today was built for one-way flow from large centralized plants to passive consumers, and it is now being asked to handle volatile renewable supply, an electrified demand that’s still scaling, and soon bidirectional flow at scale.

Smart meters and dynamic pricing are starting to expose market signals, but they are the equivalent of building the mobile app on top of an infrastructure built for the past. The main problem is that the current model assumes that the customer adapts to the system’s limitations, and not vice versa. The customer, in theory, can shift consumption away from expensive hours and toward the cheap ones. While this works well for the so-called “flexible loads”, those loads that are time-insensitive at the level of human experience (running the dishwasher at night for example), it breaks when dealing with loads that are presence-bound. Certain loads, like taking a shower or cooking dinner, cannot be shifted by any price signal, because it would impact customers’ comfort schedule.

A battery on the wall does to energy what a smartphone in the pocket did to banking.

Batteries can be charged when energy is cheap, and discharged when the customer needs it - whatever the spot price is at that time. For the first time, the household can decouple its comfort schedule from electricity price.

As mobile increased the banking events, batteries will increase energy events. Current energy providers are not equipped to read customers’ real-time signals and act on them: a battery on the wall transforms any small household energy-related decision into a data point to learn from and optimize for. The volume of energy events per customer per day goes up by orders of magnitude, and with the volume comes data, and with data comes a dominant position.

For this reason, owning only the battery on the wall will hardly make it. It’s only by owning the hardware, the software that runs it, and the retail relationship that you get to sit at the substrate of every energy decision a household makes. Every electron that flows through the battery is an intent that the system can learn from. The energy substrate can serve as demand routing for the decisions of the household, as Revolut routes our financial decisions towards its products.

But batteries can do more than that, as learning the energy needs of a house can be even more powerful than learning those of a household: knowing that the fridge is aging and its compressor is starting to suffer, or that there is one person more in the house now and the heat pump can’t warm the water fast enough, can serve as a demand origination layer. This is the most powerful layer, and it is exactly what Revolut has tried to build with in-app shopping through the RevPoints.

The cross-sell catalog for an energy company can include anything that plugs into your house or fuels it: heat pumps, EVs, EV chargers, solar additions, battery upgrades, insulation, smart appliances, even home insurance. While sales cycles are longer than for Revolut, the ACV is much higher and has higher stickiness.

To make the business case even more compelling, batteries can help the distribution grid in a way that Revolut could never do with the payment rails it uses. The electricity grid is a capacity-constrained infrastructure that pays market participants to stabilize it. A fleet of distributed batteries can participate in that market, and unlock an additional revenue stream.

This is not only good for the business, but it’s also good for the community. Europe already has roughly 110 million grid connection points, each pulling spiky loads for a few hours a day and sitting idle the rest of the time. The generic answer to this problem has so far been more grid, more capacity, more plants. However, the cheapest answer to this problem is actually better utilization of what already exists. By having household batteries participate in stabilizing the volatility, we could raise grids’ load factor from the ~55-60% it sits at today, deferring the transmission and distribution upgrades that utilities can no longer build at speed.

Unfortunately, building such a company is not an easy feat. For it to be successful, this kind of company needs three things.

First, a clear wedge to earn trust. For Revolut, this was frictionless money transfer and FX rates. For energy, the best candidates are resilience (outage backup), savings (price arbitrage, charging when cheap and discharging when expensive), synergy with solar, or a combination of all above, depending on geography. One thing to notice here is that Revolut’s customers have zero friction to get started, because all it takes is to download the app, while for energy the upfront investment needed could make the wedge less effective. For such a company to be successful, it needs to be able to “insulate” the customer from the capital risk of installing the battery.

Then, a data layer that originates demand. That’s why it is essential to own both hardware and software. A custom Battery Management System can give a level of granularity in data that no hardware on the market offers. It is important that the team treats this as engineering infrastructure from day one, because with hardware it’s impossible to fix post-purchase. As Tesla gathers data from every single car it has ever sold, the same must happen for the Revolut of Energy.

Finally, the glue that keeps everything together: sitting at the substrate is worth very little if you are not able to install what is needed to generate data and serve the customers. The winners here will either have in-house install crews (à la Base Power), or deep partnership with regional electricians. In any case, it is only with scale that comes a dominant position, and removing as many blockers to that scale as possible is extremely important. That’s why they need to deploy as fast as they can.

As I said, not easy - but a few are trying. 1Komma5º and Fuse in Europe (together with Enpal, Octopus, and a few others), Base Power (and maybe soon Tesla?) in the United States are the notable ones that have taken this approximate playbook and are doing their best to prove me right.

But here lies another big difference that makes building the Revolut of Money easier than the Revolut of Energy: scaling across geographies. While there have always been harmonized regulations in Europe and between the States regarding banking and payments, the same can’t be said about energy. It’s no coincidence that Base Power operates in Texas, Fuse in the UK, 1Komma5º in Germany, and none of them have started scaling massively outside their home market. This is because Hardware is Hard, different regulations are a painful problem to deal with, and the capital required to put a battery on every wall doesn’t move at fintech speed. An energy operator has to navigate every national regulator, every DSO tariff methodology, every aggregator license framework, every local install-and-permit code, and there is no shortcut to that.

That’s why there is the possibility that the Revolut of Energy will not be a single, massive company that scales across geographies, but rather one that grows through acquisitions, that finances itself with debt in addition to equity, and that converges on the same architecture starting from different points in different countries. The fact that it’s hard is the moat, and scaling country-by-country is the early-stage opportunity that makes this thesis relevant although the frontrunners mentioned above already have a head start over new competitors.

The Revolut of Energy is a bet on the shape of the thing, not a prediction that any single name on that list wins. Revolut got to be the substrate of money because it collapsed the friction first. Whoever does the same for energy will get there the hard way, one wall and one country at a time.

And by the time it’s obvious they’ve won, it’ll be far too late to catch them.

Subscribe now

I’m Francesco, and I invest in the backbone of the future with Visionaries Tomorrow. If you are starting the Revolut of Energy, or something as important, hit me up!

“The reactionary forces have capable men and leaders, trained to govern and who will fight relentlessly to maintain their supremacy. […]

They will try to play on the restoration of the nation state. This will allow them, in turn, to play on the most widespread feeling among the population, a feeling so damaged by recent events and so easily manipulated to reactionary ends: patriotism. In this way they can even hope to confuse their opponents’ minds more easily, since the only political experience to date for the popular masses has been within the national context and, therefore, it is relatively easy to channel both them and their more short-sighted leaders into the reconstruction of the states destroyed in the storm. If they succeed in this, the forces of reaction will have won. […]

Once the demarcation of the Old Continent is superseded, and all of humanity is united in one common design, it will become increasingly obvious that the European Federation is the only conceivable guarantee for peaceful cooperation in American-Asian relations.”

As contemporary as it can sound, the Ventotene Manifesto was written by Altiero Spinelli, Ernesto Rossi and Eugenio Colorni in 1941, during their period of internment for opposing the Fascist regime.

The authors of this manifesto were politicians and idealists, but also pragmatic enough to understand something fundamental about the modern world: scale is all you need.

Scale is the driving force of the modern world. Capitalism is built on it. Modern supply chains have been shaped around it. Scale is the reason behind specialization, it is what makes some economies more attractive than others and what makes some countries more powerful than others.

But why is that? Does it really need to be that way, or is it just something that we became so used to that we stopped questioning it?

That’s certainly what some people in power today believe, pushing for a Europe of the States as opposed to a unified European Federation. After a long period of equilibrium, things are breaking, and the reactionary forces have started again to fight relentlessly to maintain their supremacy. If they succeed in convincing some short-sighted leaders that a fragmented Europe is what Europeans need, the forces of reaction will have won.

This is why it is so important today to fight for a unified Europe. Because scale is all we need.

Most of what is possible today and was not possible in our fathers’ world is thanks to scale.

If we can buy smartphones for a few hundred bucks, if we can cover our roofs with solar panels, if we can drive 1000 KM in an EV, if we can ask our friendly AI model to spin up a website in minutes, if the banking system doesn’t feel like Venice in the 1500s anymore, if a startup can book a ride to the LEO as if it were an Uber drive, if we can print nanometer-sized structures on chips made of sand, it’s only thanks to scale.

There are several reasons why scale is such a magic enabler.

For one, scale enables depth. Deep capital markets move money to where it most efficiently creates value, aligning incentives between value creation and personal ambitions. Deep demand makes long-term investment viable and financially sustainable. This is why R&D spending and infrastructure buildout can exist without immediate financial payback. In this sense, scale enables scale.

For another, scale sets standards. Modern manufacturing is built on standardized components, making a global-scale supply chain possible, decreasing cost, and increasing safety. We don’t need to worry about our toaster catching fire anymore, because scale made standards and regulations necessary, and standards and regulations save lives.

But scale is not only about making things cheaper, faster, and better. Scale is a balancing tool that creates stability and peace, much like mutually assured destruction prevented nuclear apocalypse.

Scale, in fact, also creates leverage and optionality.

Leverage strengthens bargaining power in markets, procurement, and geopolitics. Big dogs will bark at you but won’t bite, because your decisions move such a mass that they can’t afford to stop with their own strength.

Optionality means that your system can choose to wait, absorb shocks, and decide when to act. Optionality is what moves the needle in negotiations, and thus is a game-changer both in business and politics.

The best way to achieve scale is through specialization, and in an interconnected world different kinds of specializations don’t need to happen in the same place. That’s how we ended up with the US being the experts in scale in financial systems and services, and China being the experts in scale in manufacturing and raw material processing, among other things.

The defining component of this kind of global scale is the world “global”, a free and interconnected world where everyone plays their own role without leveraging the fragility of the system for their own profit. That’s not the kind of world we live in anymore, and that’s why we, the Europeans, need to look for scale somewhere else. We need to find scale within our borders.

Where do we start, if not from fighting fragmentation?

How can 27 different nations with their rules, laws, capital markets, governments, armies, declining populations, and geographical constraints compete with the scale other parts of the world have achieved in each one of the above-mentioned sectors?

How can we not see that we need continental-scale capital markets for our companies to compete, that we need a continental-scale economy for R&D to make a difference, that we need continental-scale inertia to withstand the impact of punches from the big boys, that we need continental-scale momentum to push back and keep our seat at the table?

Europe cannot choose otherwise. Scale is all we need.

Subscribe now

My buddy Chris edited this edition of Unbeaten Path after reading The FT’s deep dive titled The Trump Supremacy. Well worth the subscription alone.

The world we live in is the best that humanity has ever had - efficient, integrated, safe and cheap.

Yet the very same dynamics that made it so good, expose it to fragility. A world in which every individual can release a software product with less than $100 but half of the internet breaks if AWS US-East-1 goes down, in which EVs have become the best possible way to move around but whose supply is dictated by China.

The beauty of all of this is that both positive and negative aspects of today’s world are unplanned effects of free and interconnected markets.

However, the times they are a-changin’ and the single points of failure that we tolerated so far can’t be tolerated anymore.

I think it’s worth stopping for a second, and exploring how and why we got here and how we can go towards a better future.

The first step is understanding how markets create this fragility.

The way I rationalized this is by drawing a parallel with social networks.

“If you’re not paying for the product, you are the product” - we know this well by now, but it’s still telling. Social networks didn’t get to the current business model because someone imposed it, but because we let the market evolve towards its optimal state.

Markets always try to optimize a (often non-explicit) objective function. Participants in the market test different variables with their product positioning, until customers reveal their preference with their actions.

In the case of social networks, many different monetization strategies were tried and the “free to use + ads” model won. What participants in the market have signaled by subscribing en-masse to Facebook & co. is that in their objective function the variable “price” had a very high weight, while the variable “privacy” was not really present.

The interesting part is that nobody would take the explicit decision to sell their personal data, but they end up doing it when the transaction is presented in a different form, behind a cheaper and better product.

Something similar has happened to the world economy over the past 50 years. Not a single country would put a price tag on its sovereignty and resilience, but as a matter of fact the participants in this game voted with their action and decided to release a slice of sovereignty and resilience in exchange of a better world.

Because, just to be clear, I really do believe that an open, efficient, specialized world is a better one. Or at least just as long as everybody agrees that it’s not worth exploiting the flaws coming with it, leaving everybody worse off.

Unfortunately, we got to a point in which these flaws are being exploited, and the fragility has become unbearable.

As with social networks, the way the world works today is the result of years of optimizing non-explicit objective functions. Resilience is hard to factor in, while costs/profits/quality are easy to optimize for and thus this is the direction the world evolved into. Markets always tend to optimize what they can measure.

The defining technologies of the 21st century are extremely complex and rely on specialization and scale to increase quality and decrease costs. This is clear looking at today’s complex supply chains, as the one of semiconductors for example, that are hyper-concentrated and have near-monopolies at every step.

This was a good deal for the West as long as the East (cough…China…cough) was OK producing low-margin peripheral goods that were serving the lower end of the market. This obviously doesn’t hold true anymore, because nobody wants to play at the edge of a large market when they can see a way to eat a bigger slice of the pie.

If you look closely, this is nothing else than Christensen’s Disruptive Innovation applied to countries instead of companies.

“Entrants that prove disruptive begin by successfully targeting those overlooked segments, gaining a foothold by delivering more-suitable functionality—frequently at a lower price. Incumbents, chasing higher profitability in more-demanding segments, tend not to respond vigorously. Entrants then move upmarket, delivering the performance that incumbents’ mainstream customers require, while preserving the advantages that drove their early success. When mainstream customers start adopting the entrants’ offerings in volume, disruption has occurred.”

What is disruptive innovation?

I know it sounds cliché for a VC, but I’m constantly amazed by this theory and by how well it explains how incumbents lose their dominant position not only in business, but also in many other unrelated fields - think of those football teams that snag up all those youngsters big teams wouldn’t let play and end up winning against all the odds.

(This is, by the way, also how Union Square Ventures became one of the best venture firms in history: investing at the edge of large markets)

If this weird mix of Facebook and disruptive innovation can explain how we got here, it probably can also explain how Europe (and more in general, the West) could stop the decline and fix their fragilities.

The first, straightforward strategy is to brute-force some changes in the cost function. This is essentially what tariffs do: they re-price resilience by penalizing certain geographies. There is plenty of literature discussing why tariffs are a good/bad means to this end, so I won’t add anything on this.

The second strategy is catalytic. Governments can decide to become the first or the best customers for the industries that matter to resilience, thus obtaining a similar end goal as tariffs. Again, this is nothing new:

“The initial growth of the chip industry from the 1950s through the early 1970s was heavily dependent on a DoD willing to pay top dollar for bleeding-edge chip technology that would offset Soviet numerical advantages in weapons systems production”

How Intel’s Innovation Problem Became a National Security Crisis

Tariffs are coercive instruments, pushing the market. Public procurement, on the other hand, pulls the market. Probably a mix of both instruments is needed, as they both tend to have a pretty short feedback loop and work well to give incumbents a new direction.

However, we need to remember that Europe today is not where China was at the end of the 1970s - we still play a role, we still have positions to defend and skills to leverage. Not all our incumbents have lost their position in the market.

Luckily, Disruptive Innovation can come to the rescue.

One of the counter-intuitive ideas Christensen introduced in his seminal work is the concept of being too close to your customers, thus missing the emergence of new needs that end up taking over industries.

Since taxes are essentially the only revenues for a country, we could define tax-payers - both corporates and private individuals - as a country’s main customers. This is why countries fail to see the emergence of new needs when these needs don’t match the needs of their established customers. If you are the US, and your customers make a lot of money selling fossil fuels, you won’t understand the needs of those that consider electricity the best form of energy.

According to Christensen, the way to get out of this dilemma is the following:

“Senior executives must first be able to spot the technologies that seem to fall into this category. Next, to commercialize and develop the new technologies, managers must protect them from the processes and incentives that are geared to serving established customers. And the only way to protect them is to create organizations that are completely independent from the mainstream business.”

Disruptive technologies: catching the wave

What this means for Europe is that governments should employ a barbell strategy: on one side, keep supporting your best customers, the industries that still make Europe great, on the other create an environment in which you let experimentation thrive, even though your main customers don’t see an immediate need.

That is what startups are for, and this is why it is so important to nurture the venture ecosystem and to remove every barrier that introduces artificial friction.

I find it fascinating that one of the best venture firms, A16Z, was founded by two among the first entrepreneurs that have focused on the internet - a classical example of disruptive innovation.

I believe this shows how comfortable with weirdness and how allergic to consensus you need to be to do the job of the investor well. It also proves that, if countries want to catch the wave of disruptive technologies, they need to create an environment for these weirdos to thrive - imagine what the USA would be today if it wasn’t for the internet and the value it helped create.

Many more internets will come, and we should create the best playground possible for the disruptors to do their thing.

Subscribe now

If you are an early-stage investor today, your worst enemy is A16Z.

You look back at the good old days when venture was a cottage industry, when the grass was greener, funds smaller and DPIs bigger, and you wish you could just go back in time before the evil gigafunds ruined everything.

They are capital agglomerators, not venture investors: they are in this game for the 2%, not the 20%. Their (venture) arrogance is so big that they believe they can generate hundreds of billions in exits, while truth is they are using this narrative to get rich from management fees while destroying what used to be a beautiful artisanal craft - early-stage investing.

You hate to be publishing another piece that opens with their name in the first line, but no one was ever fired from talking about Andreessen Horowitz, so you have to reluctantly play the game.

Latest episode that gave early-stage investors material to put into their newsletter came from Martin Casado’s twitter, one of A16Z’s GPs, when he stated that “Follow on capital tends to be more and more consensus aligned.”

The problem, critics say, is the concentration of money in the hands of a few funds like his, so big that they shouldn’t be considered venture funds anymore, that have an irresistible gravitational pull towards their ideas and thus flatten the variety of topics that get funded.

Looking at the numbers, that show an unprecedented concentration of money in the hands of a few funds and companies, some of the critics have suggested we should separate the rounds done by the gigafunds from all the rest - remember, they are not venture funds anymore, so why shall we pollute our own numbers with theirs?

I believe this is a dangerous argument we are getting into.

It’s risky because saying “they’re not a venture fund” just kicks the can down the road without addressing the core of the issue: venture investing has changed at every stage, and if you look closely you can see the same dynamics happening also within the early-stage investors community, even in those deals that are not yet “polluted” by gigafunds’ arrogance.

The idea of “non-consensus investing” has always been appealing for VCs, and it makes a lot of theoretical sense: to capture the most out of a transaction you want to find deals that have the biggest delta between perceived value and real value.

The issue is in the word “perceived”.

Venture Capital is a weird industry, that needs to generate returns 10 years in the future, but that assesses quality looking at the past. There is no real way to define a Tier 1 fund, except by looking at what deals they’ve done in the past inferring their multiple at exit, or looking at who has invested in companies that are (now, in the present) perceived as winners.

It feels wrong, I know, but even the most die-hard haters of gigafunds can’t avoid celebrating big rounds of their portfolio companies at high valuations as if that was the end goal of building a company. Also these people need to raise next fund, have LPs they need to pitch to and need something to show.

Today’s world of venture is one of storytelling and hype, one of concentration and kingmakers. In this world, being contrarian is riskier than ever.

If you believe that this happens only when gigafunds are involved, you are very wrong.

Rounds are getting bigger, at every stage, and are getting smaller in number. This means that, even if you play at the earliest stages, you can’t avoid to consider what your downstream funds like. That is true even if the downstream funds are not the gigafunds everybody hates.

This feels wrong, and I bet (even though I don’t have the data to back my bet) that this is not serving well the overall ecosystem. When only a few topics get funded, when a few kinds of founders get backed, we are implicitly capping the overall value we are creating for the society.

However, as wrong as this all feels, venture funds need to survive - need to be top of mind for founders and raise money from LPs.

It is proven that it’s not worth trying to time bubbles in public markets. Even if you think the markets are overvalued, you are usually better off riding the bubble until it bursts than getting out of the market and waiting they return to fair valuations. In a market that grows out of control, the opportunity cost of not riding it is very high.

I have a feeling we are living a similar moment in private markets. You might think the next AI coding company will be out of business in the next few years, but when these companies raise a round every 6 months doubling their valuation each time, you’re probably better off investing in them anyway.

I know, public markets are liquid so you can get out fast when everything starts collapsing, while you can’t do the same with startups.

But “markets can remain irrational longer than you can remain solvent” or, in VC words, your bet can stay contrarian longer than the company’s runway. And in a bubble market like the one described above, contrarian bets are crazy scary.

Can you really blame early-stage VC for playing the game?

Subscribe now

My buddy Chris edited this piece while rage-clicking through Runway.ai. After three hours, he gave up on making a video after realizing he’d need a PhD in Promptology and Machine Whispering.

Paul Graham is, in my opinion, the best storyteller of the startup world.

He is the one that coined the most famous line of this industry: “Make something people want”

If you want to understand that line, you need to read his 2004 essay titled “How to make wealth” - probably the first essay in which he wrote that line (but don't quote me on this).

Reading something that was written 21 years ago, talking about such a fast changing world as the one of technology, might seem dumb but I promise it's not. It actually helps framing the current state of venture much better than most of the things I've read recently.

One of the paragraphs is titled “Technology = Leverage” and within that, Graham claims that the best way for a startup to be successful is to pick a hard technological problem.

What is technology? It's technique. It's the way we all do things. And when you discover a new way to do things, its value is multiplied by all the people who use it. It is the proverbial fishing rod, rather than the fish. That's the difference between a startup and a restaurant or a barber shop.

It's fascinating to see how to the eyes of 2004 Graham, the essence of a startup was to create new technology - the harder, the better. Programmers for him were artisans, that could literally create wealth with their hands by writing pieces of software.

I hope that by reading that essay in 2025, you too see the irony of having companies worth dozens of billions that are nothing more than wrappers around someone else's technology, indistinguishable under the hood one from another.

How is that possible? What happened to the “discover a new way to do things”? What happens to the “make something people want” when the “making” part becomes so easy that you can literally rebuild popular websites with a few prompts?

Don't get me wrong, Lovable and co. are exceptional companies and I only have admiration for the people that built them. Moreover, from a purely Economy 101 point of view, these companies are exceptional at making something people want, because people are paying their hard-earned dollars for what they make.

However, this new world makes early-stage investing extremely hard.

Early-stage investing has always been mostly about people, but reading people has fundamentally changed.

In a world where making things was actually hard, an early working product or relevant experience was a good predictor for the success of a startup. You could get a sense of the quality of the founder looking at the early product.

Unfortunately, we don't live in that world anymore. What is it that we should look for in founders at this stage then?

If building a product is now easy, all we have left is the ability to sell that product. The ability to build a compelling story around it.

When products are almost indistinguishable from each other, the greatest value lies in always being top of mind, at all costs.

Entrepreneurship at the early stages is almost 100% convincing others - investors, employees, partners - so it shouldn’t be a surprise that storytelling is number one in the list of must-haves.

The fine line between storytelling and bullshitting

However we need to be very careful here, because storytelling is very different than bullshitting, but the line between the two is very fine.

How do we trace that line?

The two traits that help defining it are, in my opinion, consciousness and authenticity.

Storytelling without consciousness is wishful thinking.

Consciousness means knowing where someone comes from, what surrounds him, and the difficulties that will be encountered during the journey. While every single founder, looking back, will say the journey was harder than imagined, having an idea of the possible obstacles during the way is important.

Consciousness of the surroundings is also essential when taking strategic decisions - the steps to build something valuable are different if you are building in Texas, in Germany or in Singapore. Sometime this trait can also be described as “self-awareness”.

Authenticity also needs to go together with storytelling.

Storytelling sometimes pushes people to paint themselves like someone they are not. This is very common when fundraising: you can’t fake forever, but you can probably play the role for the limited amount of time you spend with an investor during the due diligence.

Someone who is authentic is often transparent, reliable and consistent in the behaviors - all traits that are good when doing business.

What makes evaluating these traits particularly hard, is the limited amount of time in which we are supposed to make these decisions.

For example, even though there is some merit in evaluating the personal background of founders, it is extremely hard to efficiently factor it in the decision making process.

The time you get to spend with the founder before the investment is simply not enough, especially if the founder is building something valuable and other people know it. Moreover, there will always be some role playing during the due diligence process, and the founder will hardly expose to an investor their personal struggles or stories about their past.

Professional background might be easier to evaluate, but being a public and easy benchmark, it will be certainly already priced in.

There is still some room for reasoning here: is the professional background an indicator of success? It might be, but repeated founders certainly put a premium on the price of the round.

The problem I see with repeated founders is that it might be hard for them to deviate from what they have done successfully in the past, which might cause some version of innovator dilemma to kick in. So I believe that, while relevant professional background is certainly needed for some technical roles, it is not always worth the price premium it puts on the round.

A word of advice

When reading personal opinions like the one in this piece, you should always be careful: if it was easy to write a playbook to invest at pre-seed, VC wouldn’t be such a risky asset class. Also, if I held the secret to successfully picking early-stage founders, I wouldn’t be here writing this, but I would be sipping a cold beer on my private island instead.

If there’s one thing to take away from this piece, it’s this: embrace storytelling, reject bullshit.

My friendly Roman Francesco edited this edition and helped turn a rant into, well… you decide!

With every fresh declaration, I couldn’t help thinking that it sounded like complete and total bullshit. Every update was met as if it was the second coming of Christ. It’s been going on for just over 2.5 years and while the technology has come on leaps and bounds, it is still limited and the improvements, to me at least, were becoming less profound.

I’m of course talking about AI. For the past six months or so, OpenAI’s executives have been talking about artificial general intelligence as if it were literally around the corner. An extremely near-term inevitability.

Sam Altman himself described GPT-5 as “a significant step along the path to AGI”. As an ex-tech journalist and PR pro, OpenAI’s PR team executed the game plan almost to perfection. Hints, leaks, and carefully stoked expectations painted a picture of GPT-5 as the moment the company would cross some invisible line between powerful tool and proto-AGI.

The campaign only fell down in one place. Unfortunately, it tripped up at the most important part. The actual product itself. And the problem with the kind of hype narrative stirred up by OpenAI is very simple: if you don’t deliver, you disappoint.

Benchmarking Reality Check

For balance, it is better. It’s faster and smarter. For example, it scored 74.9% on SWE-bench Verified compared with GPT-4.1’s 54.6%, according to data from Leucopsis. Cornell University researchers found that in medical reasoning, it improved reasoning and understanding by 29.6% and 36.2% respectively over GPT-4o.

GPT-5 even shows promise in mammography image interpretation, however it still lags trained clinicians in some areas of that field.

Don’t get me wrong, this is all great stuff but for all of GPT-5’s polish (and there’s plenty), it failed to meet expectations. It’s not a leap towards AGI. It’s barely more than a pigeon step towards AGI. Don’t take my word for it.

“GPT‑5 failed the hype test,” writes The Verge, noting that despite grand promises, users reported only incremental improvements and found the model “underwhelming” and “not a massive step forward.”

It’s incremental, not transformational. The narrative was akin to landing humans on Mars, but instead we’ve got a software patch. A capable upgrade and nothing more.

The Inverse Exponential: Fireworks First, Sparks Later

Privately, I’ve long argued that AI’s trajectory is wrong. The narrative has suggested endless exponential gains, yet reality has been the exact inverse. The big leaps were early and now are being followed by smaller hops or even nudges.

Like fireworks, the biggest, brightest bursts come first, followed by smaller pops that still light the sky but no longer dazzle in the same way.

GPT-5 is another spark, not the grand finale. Mobile phones weren’t indispensable overnight. They became mainstream through hundreds of tiny and boring iterations by Swedish, Finnish and Canadian companies. Even when smartphones arrived, they exploded on the scene and the improvements like better batteries, sharper cameras and faster chips aren’t lifechanging.

AI may now be entering that same phase.

The Cost of Over-Storytelling

The danger of overselling is that it erodes credibility. If every new release is billed as “a step toward AGI” and you don’t deliver, the credibility gap widens. Gartner calls this the “cycle of cynicism.” Once customers, regulators, and investors feel burned, they stop evaluating each claim on its merits and start assuming all claims are inflated.

The problem with the credibility gap is that it has a long tail. In consumer markets, customers churn by declining upgrades or downgrading to free tiers. In B2B, execs set tougher benchmarks or ask for longer pilot projects. In capital markets, it kills valuations or, worse, leads to ferocious corrections when everyone agrees that it’s overhyped.

For OpenAI and its peers, that’s not just a PR stumble. It’s a strategic risk. Trust is far harder to rebuild than it is to maintain. Microsoft learned this in the 1990s after over-promising on early Windows launches. Tesla has learned it repeatedly in autonomous driving claims. It’s even worse for the leading LLMs. Realistically, 99% of use cases are exactly the same, so OpenAi, Anthropic et al won’t generate blind brand loyalty (like Apple), so they don’t have a die-hard fanbase (unlike Tesla).

Ironically enough, the quantum industry has long been accused of being overhyped and every quantum company I’ve ever met is at pains to be as factual and sober as possible. I spend a lot of time advising quantum startups to show off a little ankle and flirt more.

The Investment Paradox

Yet in AI, the money keeps flowing. In the first half of 2025, AI startups pulled in 53% of global venture capital funding, and 64% in the U.S., according to Axios Pro Rata. More than half of every VC dollar on earth is now chasing AI. It’s distortion.

OpenAI itself raised $40 billion at a $300 billion valuation this spring and yet, Sam Altman, architect of the hype cycle, has been unusually candid: “AI is in a bubble… someone is going to lose a phenomenal amount of money”.

This is the paradox: investment is peaking just as narrative credibility is slipping. The checks are still being written, but the stories behind them feel shakier. You don’t often see this much capital sprinting into a sector just as the sector’s leading figures admit the hype might be outpacing the fundamentals.

It feels eerily like the dot-com bubble: money flowing not despite skepticism, but because investors fear missing the next Amazon. However, back then the storytelling hype was far less enthusiastic.

The Real Lesson: Storytelling Discipline

So what’s the takeaway? AI isn’t entering the trough because of its actual capabilities. They are real, and improving. It’s entering the trough because of the way those capabilities were framed. GPT-5 wasn’t bad. Far from it. It was incremental, dependable progress — arguably the kind of progress enterprises need most. But the story told was bigger than the product could carry, and the fallout is that trust takes the hit.

This is the lesson for founders, investors, and technologists: storytelling is leverage, but it’s also liability. Hype can bend markets and pull forward adoption, but if it overshoots, it plants the seeds of disillusionment. And when storytelling outruns reality, reality always wins.

Better to promise steady progress and surprise people than to promise AGI and deliver a service pack.

Gartner’s Hype Cycle: Sliding Into the Trough

This is why Gartner’s framework resonates so strongly right now. The 2025 Hype Cycle places Generative AI squarely in the ‘trough of disillusionment’, according to Gartner. Even The Economist put it bluntly: “Many bosses are sliding into the ‘trough of disillusionment’”.

The irony is that GPT-5’s incrementalism is exactly what we should have expected, which is a tool getting steadily better, a platform refining itself. A trajectory toward usefulness rather than transcendence. But because the story was overplayed, the trough feels deeper.

When Hype Meets History

And this isn’t the first time we’ve seen the script. Virtual reality was going to replace the office. Blockchain was going to replace the bank. Autonomous cars were going to replace the driver. Each promised a revolution; each hit the wall of complexity.

The pattern is eerily consistent: the initial leap dazzles, early adopters rave, capital floods in — and then the technology stalls on the hard edges of physics, cost, or usability. AI isn’t immune. GPT-5 doesn’t mark failure, but it does mark normalization. The shine wears off, the story gets less sexy, and the real work begins.

The Plateau of Productivity Is the Prize

The hype cycle doesn’t end in despair. After the trough comes the “plateau of productivity”, which is the point where technology matures, becomes boring, and quietly powers the world. Cloud computing is the obvious example: no longer hyped, but utterly indispensable.

GPT-5 may be less a letdown than the beginning of that plateau. A shift from dazzling demos to dependable tools. The irony is that the plateau is often where the real money is made. AWS wasn’t built at the hype peak of the cloud; it thrived when the hype was gone and the boring work began. GPT-5 might represent that same inflection point for AI.

Where Next?

The good news is that this doesn’t spell collapse. It spells maturity. If the exponential dream was always a myth, then what’s left is the durable, steady growth of a technology finding its real place.

GPT-5 isn’t AGI. It’s something more pragmatic: a plateau of productivity, where AI is less about headlines and more about infrastructure.

And maybe that’s the real story worth telling.

I can’t help thinking about this proverb when I look at what’s happening recently between the US and China, thinking about the opportunity that lies ahead of us as Europeans.

This is a long post, but one that means a lot to me and should also mean a lot to you if you’re European.

Enjoy!

My buddy Chris edited this edition of Unbeaten Path after closely following his friend's (and ex-colleague's) coverage of the MAGA meltdown over Jeffrey Epstein. You can sign up for Tara Palmeri's Substack here.

Europe has a huge storytelling problem.

Some might argue storytelling doesn’t matter because there’s a difference between what people say and reality. However, this is the short-attention-span era: we've outsourced critical thinking to algorithms, owned by the US or China, and therefore we simply tend to believe what people claim is the current state of things.

I see this dynamic every day in my VC job, and I’m seeing it play out in the latest iteration of the US vs China debate.

The fact that I’m referring to it as “US vs China” says it all: where the hell is Europe in all of this?

I’ve read dozens of articles and listened to even more podcasts that discuss the levers the US should pull to regain a position of strength or analyze why China will be the dominant power of the next century.

Unfortunately, I haven’t come across a single piece discussing what Europe’s role is in this and how Europe could regain strategic relevance in this fight between the two giants.

So, I decided to write that piece.

What follows is my attempt to paint an optimistic picture – but always grounded on truth - of what I believe is the state of things and why Europe has a chance to play to its strengths to regain relevance.

The US needs China

The first thing to notice, which not many American opinionists like to admit, is that the US needs China.

What's very clear by now is that the US has become an economy driven by financial dominance, and it has 36 trillion reasons why this can't change.

The modern US has been defined by the dominance of the dollar as the exchange currency of choice for global trade. This has generated infinite demand for US treasury bonds, which in turn provided never-ending fuel to internal consumption.

What happens when the well-oiled system of global trade encounters additional friction? Markets know the answer very well and reacted accordingly when Trump imposed his tariffs.

The US needs a trade deficit because they need to release US dollars for countries to buy their debt. This is the case for most countries and especially so when talking about China as the US accounts for around 15% of Chinese exports. How do you replace that?

Moreover, most of the value US companies create today is through their contribution at the top of the value chain – their IP, their designs, and their distribution networks.

They need a manufacturing partner to put their IP in the hands of their end customers. Hello “designed in Cupertino, manufactured in China”.

The way global trade is organized today works very well for American companies: Nvidia and Apple wouldn’t be worth what they’re worth if it wasn’t for the extreme efficiency of specialized manufacturing in places such as China and Taiwan. Big tech can afford to spend $100Ms to poach the best brains from the competition because global trade allows them to monetize on the IP those brains will generate.

The Trump administration wants to increase the cost of producing abroad by imposing tariffs but, as discussed above, you can’t have your cake and eat it too – increasing friction on global trade increases the cost to finance their national debt, which by now is so big that it can bankrupt the country.

To add complexity, China controls the supply chain of many of the basic raw materials that define the technologies of the future. As much as Americans try to play hardball, they know they can’t get around many of those monopolies overnight. China obviously knows that as well and has already used this weapon in the latest negotiations over tariffs.

While it’s not easy to say if and how the US will be able to decouple from China, it’s clear that it will take time and be very painful.

China needs the US

From all of this it might sound like China has already won, and it’s tempting to think so, but that’s not the full reality of things either.

In fact, China needs the US, too.

People in the Western world are beginning to experience the benefits of moving up the value chain, as they drive Chinese cars, purchase Chinese smartphones, and record videos using Chinese drones.

We shouldn’t be surprised that we are starting to see Chinese high-tech products entering our lives, as it is the result of a shift in public spending from real estate to industry after the real estate crisis of 2020.

When companies are paid to produce more, they will produce more even if that means producing more than their market can absorb. This creates downward pressure on prices and saturates internal markets – without export, this system collapses.

Without the relief valve of export, Chinese producers will have to look even more at the internal market. This is very risky for the Chinese economy, because too much supply and not enough demand create deflation, which is the nightmare of economists because there are no effective economic measures to defeat it.

We are used to thinking that the Chinese economy is mostly focused on export - and they export a lot, of course – but what remains often untold is that, as a percentage of their GDP, they don’t export that much.

They are a big country and don’t really export that much compared to what they produce. Despite this, the rest of the world feels overwhelmed by Chinese goods.

This is why they need the US: they need someone to sell their goods to.

China obviously knows this, and in the last years has tried diversifying their export markets with programs such as the Belt and Road Initiative and the investments in Africa. But looking at GDP per capita in the world, it’s clear that the US are hard to replace:

Who’s going to spend as much as Americans do? Hardly anyone.

Arrived at this point, it is clear how complex and high-stakes the current US vs China debate is and why those trying to simplify it into slogans are not painting the full picture.

Europe needs both the US and China. Or does it?

So, where is Europe in all of this?

We are starting from a weak position, and it’s all because of fragmentation.

Fragmented capital markets don’t have enough scale to support big companies and to attract foreign investments; fragmented bureaucracy creates internal barriers that look very much like tariffs, and fragmented defense makes us spend money slowly and inefficiently when what we need is strength and deterrence.

This means that we can’t attempt financial dominance like the US, we can’t grow European champions that can win internationally like in China, and we can’t impose our military power like the US and China.

Moreover, we are not energy independent, thus being susceptible to external shocks that we can’t control - such as a crazy guy invading a neighboring free country.

Looking at the glass half-full, starting from a weak position means that we don’t have a strong position to defend in this real-world Risk game. We have seen that the main challenge China and the US are facing is to solve their problems without breaking what already works. We don’t have this “problem”.

Being a first mover works great when it works, but it means you have to face uncertainties without a working playbook to rely on. The US and China have been first-movers with their own strategies, Europe is today a follower: we can look at what worked and what didn’t work and try not to commit the same errors.

The most pressing thing to fix in Europe (leaving bureaucracy aside) is energy, as “everything is downstream of energy”.

Not many know that today Europe produces more electricity from renewables than from fossil fuels.

The beauty of renewables for Europe is that we still play an important role in the value chain of wind and hydroelectric power. Moreover, while most of the solar industry is concentrated in China, it is mostly the result of the difference in investments between Europe and China - 10x less in the last decade - and there is nothing structural that cannot be reverted (apart from a few selected raw materials). This is a big difference from fossil fuels, that will always be constrained by the availability of the raw materials in a given region.

Electrifying our economies can reduce our dependency on oil-producing countries, many of which have unstable or openly hostile governments, finally making obsolete fossil fuel supremacy and achieving strategic independence. This should be a major geopolitical goal for our continent.

The other important thing to notice is that complete decoupling of modern economies is impossible to achieve, and it is delusional to think of a future in which the US or Europe or China will produce all they consume on their own.

If this is true, then what is it that we do want to reshore and own?

Things that have high value added, that employ skilled people in interesting jobs, and that are in the supply chain of the most critical industries.

As it turns out, in Europe, we still have much of that. Despite the talk of the town being how the only industry we have here is that of bureaucracy, the reality is that European countries are still leaders in the manufacturing of industrial tools that are essential for precision engineering, pharma, and semiconductors.

We still are a manufacturing-heavy economy, and we should always remember that when talking about “reindustrialization”.

Precision machinery is essential for the future of industry: a future in which energy usage comes mostly in the form of electricity, creating a natural ecosystem for software-defined manufacturing to thrive and for the virtuous data feedback loop, the loop that made internet companies great, to kick in.

Sure, software will play an increasingly important role. But at the core of everything is the ability to build high-quality physical things with sustainable marginal costs, something that is not true in the internet world and that economies can’t relearn overnight. Something that, despite everything, Europe is still quite good at doing.

One of our main weaknesses is having almost completely missed the internet revolution, as exemplified by all those memes comparing the market cap of the Mag-7 with that of entire national public markets.

But that is also a trap for the US, as those companies have built their brand on software but built their moat with hardware, often not built in the US.

It’s a familiar pattern: railroads needed land, cloud computing needed fiber, and now AI infrastructure needs vast, stable power. In the exponential age where compute demand outpaces supply, owning tomorrow’s terawatts may matter more than having the smartest model.

Azeem Azhar on Linkedin

If this is true, and if Europe is ambitious enough to remove the artificial blockers that bureaucracy and fragmentation have imposed on our economies, why should we accept a subpar role in the new world order?

Why shall we still accept to be the supporting leg of the empires of yesterday or of those of tomorrow?

I don’t think we should, and people are starting to realize that.

Subscribe now

This post was written without the reassuring supervision of my friend Chris. Be forgiving if the style isn’t as sharp as always.

In this understanding of tech dominance, the driver of generational change is a paradigm shift: from mainframes to personal computers, from desktop applications to the web, first on personal computers, and then on mobile. Each shift brought a new company to dominance, and when the next shift arrives, so will new companies rise to prominence

The end of the beginning

Many in VC are claiming AI has killed software, and that’s why it’s only worth investing in hardware today.

I don’t agree with this claim.

I believe AI is a paradigm shift for the world of bits: the way we interact with software will change because of AI, and so all the infrastructure of the web will have to change accordingly.

Let’s take the internet as an example. Human attention is the main value driver on the web - catching and selling human attention through ads is the only business model that has worked at scale, and it’s what has allowed Google and Meta to become what they are today.

All the big paradigm shifts Ben Thompson mentions in the quote above - “from mainframes to personal computers, from desktop applications to the web, first on personal computers, and then on mobile” - had to do with the way we interact with software (and the time we spend doing so).

This shift will create disruption and will force business models to change: some won’t work anymore, and some new ones will emerge.

If we start interacting with the digital world via other means that are not a screen, what happens to Apple’s app ecosystem?

If humans don’t browse websites anymore, delegating that to agents, what happens to Google’s ad revenue?

I think the concept of agentic web is fascinating, and there is a lot of value to be created. In software, certain technological innovations become paradigm shifts not because they are hard to copy, but because they enable new business models and market dynamics that compound and help create stickiness - i.e., create a moat

Many in VC are also claiming that AI is a paradigm shift for the physical world.

I don’t agree with this claim either.

In the physical world, AI is an enabler - it has the potential to make everything more efficient, more intelligent, more useful, but generally doesn’t change the incentives to buy or use a product.

For this reason, I believe it is an error to treat physical AI like a paradigm shift.

If we apply AI to existing markets, using it as a differentiating factor, we are implicitly taking a market risk: we are assuming that something that didn’t sell before AI, will start selling when we add AI.

There are some edge cases in which this might work, maybe if we are able to build a humanoid robot that does chores for us someone will want it, but I hate market risk, so I don’t buy into this vision.

Does it mean we shouldn’t be bullish on AI applied to the physical world?

Absolutely not! We should be, and I’ll explain why.

Widespread adoption of transformer-based AI models (what everyone calls “AI” these days) is a productivity shift that has come at the same time as a real paradigm shift for the physical world: the end of the Pax Americana with the geopolitical order it created, and the beginning of a transition period that has brought extreme geopolitical uncertainty.

The latest chapter in the US-China cold war is the export ban on rare earth metals. That China has a near monopoly in this is by now very clear, but why is that?

It’s not only a matter of raw materials; in fact, China imports many of the raw materials that it then processes. It’s a matter of technology, know-how, and willingness to industrialize the combination of the two. It’s not a case that technology and machinery are on the export ban list..

This has enabled China to offer a product that buyers need at a price that makes it the most wanted on the market. In this case, the product is also a commodity, and in a normal state of the world price is all that matters for commodities - easily explained the monopoly.

Will AI change this? Will transformers make rare earth metals refined in Italy, Texas, Ohio, or Germany more appealing to the market? No, unless their cost suddenly matches the willingness to pay of the customers. This brings us back to the main point: all of a sudden, certain buyers have a higher willingness to pay, and we can use the latest technologies (including AI), to supply a product that matches that demand.

This geopolitical shift changes incentives in the biggest industries out there, providing a new market and customers that have strong intentions to buy - the famous urgent buyers. This will make some business models outdated, creating disruption in otherwise crystallized industries, and allowing new entrants to undermine incumbents’ domination in these sectors.

New markets are way stronger moat than new technologies for young startups. As the world evolves and needs change, it’s easier for startups to tackle these changing needs than it is for large incumbents:

Why do incumbents immediately copy new technology when they see it is working but don’t immediately enter new markets when they see they are working? […] The answer is that to adapt they have to feel comfortable with how uncertain the new technology or new market is. Incumbents strongly dislike uncertainty so they wait for it to be mitigated. But startups can build moats in new markets while they are still uncertain where they usually can’t with new technologies.

https://reactionwheel.net/2020/11/productive-uncertainty.html

I think physical AI has the potential to change the world, but we shouldn't fall into the trap of throwing money at anything that has some intelligence in it - that would burn money and momentum, and we need both of them right now!

Subscribe now

Francesco Ricciuti edited this edition of Unbeaten Path by underlining sovereignty.

I’ll answer first: I don’t.

Judging by the grief porn littering my social media, this is clearly a controversial view but hear me out.

Eight years ago, squeezed into a Brussels phone booth, I asked Martin Peers, then working with The Information, how a scoop-hungry European tech reporter like me could ever make Europe matter to Silicon Valley.

Spoiler alert: I couldn’t. I thought the tech regulation angle would resonate but no, the Valley always assumes the EU regulates, so the detail didn’t matter.

This is exactly why The Information has succeeded. The leadership knew exactly where its strength lay: The Valley, The Valley, and… yes, still The Valley.

TechCrunch has always been predominantly a US publication. Scratch that. It’s overwhelmingly a Silicon Valley publication. Sure, Natasha, Paul, and Mike have written thousands of articles, and Romain Dillet, who helped define TechCrunch’s reputation in France, authored no fewer than 3,500 posts. But hundreds of thousands (perhaps millions) of articles have originated from a small slice of California, 5,668 miles (9,122 km) from Paris.

Underfunded Newsrooms

When I was reporting, I chased insiders because that’s where the premium in news lies. Now, as a reader, I still seek that insider’s perspective. TechCrunch’s European team did stellar work in that regard.

Anna Heim interrogated founder Michael Kissner to introduce Akhetonics, a small but powerful German photonics firm, to a global audience. Paul Sawers diligently covered open-source long before the AI and LLM wave swept in and elevated Konstantin Vinogradov’s little known OSS index into a scoreboard recognized worldwide. Mike and I have shared more pints than story pitches, underscoring the dying art of source building and shoe-leather reporting over regurgitating a press release.

But consider this: Akhetonics wasn’t just a chance for me to self-promote a portfolio company. It was one standout player among a sea of deeptech firms competing desperately for the attention of a few journalists. How can a handful of reporters write about an entire continent? Well, they can’t.

I spent a few years covering tech in southern Africa for Bloomberg. The Johannesburg bureau had around 20 journalists. We dipped into stories across the continent but the focus was on Africa’s most sophisticated economy and that reflected in the strength of the coverage.

There are around seven PR professionals for every reporter. I guess there are perhaps seventy for each TechCrunch journalist. This imbalance, to me, always seemed unsustainable. TechCrunch’s European reporters cared deeply about Europe; its management didn’t or couldn’t afford to.

Subscriptions Models and Media Economics

This isn’t unique. US media often covers the rest of the world as an odd peculiarity - a gentle reminder to American readers that, hey, there’s a big world out there. If you think I’m exaggerating, I’ll see your scepticism and raise you a curious New York Times piece lamenting London’s supposed abundance of “porridge and boiled mutton.” Right… me neither.

For balance, European media isn’t immune either. Sky News gleefully highlights bizarre stories. From US shoplifters wielding pythons or Sweden’s annual moose migration. I love this stuff! But if I’m American or Swedish, I’m not relying on Sky News for my local updates. Europe’s tech ecosystem deserves a similar self-reliance, especially as the region grows more strategically sovereign. Dedicated media coverage is part of that sovereignty.

Let’s be honest: TechCrunch pulling back isn’t exactly an existential blow to European startups. Investors aren’t suddenly operating blindfolded, nor will founders vanish overnight into obscurity. Instead, the real loss is symbolic. It’s a stark reminder that Europe has long been treated as a mere footnote by US-based publications.

Contrast this with The Information, thriving on its subscription model, boasting around 45,000 readers who willingly shell out at least $399 annually, plus punchy ad revenue from a newsletter reaching hundreds of thousands. Yet, I’ve never once been sent (or even seen anyone request) a paywalled article from The Information.

Sifted, however, is a different beast altogether. European VC WhatsApp groups buzz constantly with requests for screenshots of its paywalled articles (you know who you are). For context, my sub costs just thirty quid a month. Europe clearly values dedicated, insider-driven reporting—just not enough to universally pay for it, apparently.

Media Sovereignty

Europe is alight with sovereignty at the moment. Just this morning, I watched a podcast with Andreas Monk of EUVC on precisely that topic. Why limit sovereignty to matters of defense or next-gen computing?

Europe needs media champions that don’t merely glance across the Atlantic for validation.

Scrappy local outlets like Tech.eu (cheers again, Robin Wauters, for letting me pilfer your data back in my reporting days), EU-Startups, Tech Funding News, The Next Web and Resilience Media are already stepping up.

Actively investing in these voices by even just reading and sharing their work is how we turn Europe’s media landscape into something robust, influential, and genuinely sovereign.

TechCrunch’s chairman insists they aren’t leaving Europe, promising great things via their acquisition of Foundry. Yet, it’s hard to see how this happens after axing decades’ worth of experienced European talent. It smells suspiciously like AI-generated filler to me but more on that in a future edition of Unbeaten Path.

That said, TechCrunch remains my go-to for US tech news, ironically enough. Despite my huge admiration for the FT (via my good friend at their San Francisco bureau), the FT isn’t my number one choice for US tech updates. Though credit where it’s due: the FT recently published a terrific Marc Benioff interview that’s worth ten minutes of anyone’s time.

TechCrunch retreating from Europe? Adios. There’s enough money (and demand) for a commercial news organization to enter the fray and make the indefatigable team at Sifted sweat. They already get heat from Riddhi Kanetkar, Cate Lawrence, Lucy Adams, Siôn Geschwindt and Sofia Chesnokova, but let’s add some more.

Europe’s tech deserves its own headlines, not hand-me-downs. Let’s start building them.

My buddy Chris edited this version of Unbeaten Path having just returned from the rolling green fields of rural Ireland 🍀

Over the past weeks, variations of this graph have flooded my Linkedin.

I am not sure why people are only now realizing that China is miles ahead in electricity generation. China overtook the US in 2010… but it’s a good hook to talk about something close to my heart: there's not a single reason why we should reduce energy consumption.

In fact, if anything, it should be a primary goal of governments to be sure that citizens have access to abundant energy in the most reliable way possible, at the lowest possible price.

The reason why is pretty straightforward:

In the last decade, energy has been discussed mostly in the context of the Paris Agreement and the Net Zero 2050 plans. Unfortunately, when discussed under the climate lens, energy consumption has been demonised and consumers have been held responsible for something they didn’t really have much power over. No pun intended.

Given the huge momentum behind it, the Paris Agreement was a big opportunity for Europe to realize that its dependency on fossil fuels was extremely dangerous and for governments to invest in new ways of producing energy.

Despite leading the way in non-fossil energy production, Europe didn’t do enough and still relies too much on imported energy:

The fact is that, for a very long time, we lived in a world where relying on imports wasn't a problem. In that world, imagined in Bretton Woods after WW2 and powered by US dollars, fossil fuels were the ideal source of energy - easy to transport and to transform into electricity or work. This held true as long as the US was a net importer of fossil fuels, and thus it was in their interest to protect global trade with their military might.

This world ended in 2019 when, thanks to the Shale Revolution, the US became a net exporter of fossil fuels.

As the United States transitioned to a net oil exporter post Covid, its incentive to protect global shipping lanes waned

The New Joule Order

This was the nail on the coffin of a system that was already shaking under the pressure of the growing US debt, and that culminated in Trump’s recent approach to international relations.

As I’ve said multiple times, we are living through a geopolitical inflection point that will reshape the global economy. This is a big, big opportunity for many who want to provide new solutions to new problems.

Energy is no different. For Europe, the economic argument, the security argument, and the environmental argument are now well aligned and pushing for more local and sovereign energy production.

For a long time, green energy has suffered a green premium compared to fossil fuels. It had to account for expensive infrastructure and inefficient production. However, in recent decades, renewables have become much more affordable and fossil fuels have become more expensive because of tariffs.

Moreover, the primary source of fossil fuel for Europe, Russia, has brought war back to European soil and our friends in the US, from whom we buy a lot of gas, have become unreliable. The security argument is now stronger than ever, and it’s not by chance that most European countries are openly talking about nuclear now - something that Net Zero didn’t even go close at doing.

The interesting part of this situation is that concerns over security might end up having a bigger effect on the energy transition than concerns over climate change.

Often chaos can be a positive force of change, and in this case these three forces push in the same direction in something that looks like a win-win-win situation.

Mind you, it’s not going to be easy. There are many problems to be solved.

The first problem is that of raw materials: China has a near-monopoly of many of the materials needed for the energy transition, and has already started using it as a negotiation tool to reduce tariffs imposed by the US.

China is on its way to becoming the world’s first “electrostate”, with a growing share of its energy coming from electricity and an economy increasingly driven by clean technologies. It offers China a strategic buffer from trade decoupling and rising geopolitical tensions with the US. The country is not only rapidly advancing towards self-sufficiency in energy from secure domestic sources, but also wields vast power over the markets for the resources and materials that underpin technologies of the future.

The Financial Times

Magnets are essential for any energy production technique that implies turning movement into electricity, and if we want to step up local energy production in Europe we need to have reliable and cheap sources of the materials needed.

An additional problem is that of infrastructure. In a fossil fuel-powered world, the metric to track to understand how good an energy production means was used to be the levelized cost of energy. This is the average cost of electricity generation over the lifetime of an energy asset.

It made sense because shipping was straightforward and without losses, so produced energy was more important than delivered energy. When energy is produced directly in the form of electricity, which is very hard to store and needs to be distributed immediately at the highest efficiency possible, delivered energy becomes the metric to track.

Not only does The New Joule Order put an emphasis on security as a key driver, but it also emphasizes delivered energy to consumers in the form of joules. Delivered energy holds greater value than produced energy, which emphasizes return on equity (ROE) as opposed to achieving the lowest levelized cost of energy (LCOE)

The New Joule Order

What does this mean in practice? Infrastructure, infrastructure, infrastructure.

We need to become good at planning, authorizing, and building infrastructural projects to support local production of energy.

The final question that comes to mind is how to finance these projects.

While there is certainly a case for VCs to provide a part of the financing, at least at the earliest stages, we are talking about huge numbers here - $68T according to BlackRock’s Larry Fink:

As we enter our century’s second quarter, there's a growing mismatch between the demand for investment and the capital available from traditional sources.

Governments can’t fund infrastructure through deficits. The deficits can’t get much higher. Instead, they’ll turn to private investors.

Meanwhile, companies won’t rely solely on banks for credit. Bank lending is constrained. Instead, businesses will go to the markets. […] Throughout history, infrastructure has driven a surprising amount of economic growth. […] Today, we're standing at the edge of an opportunity so vast it's almost hard to grasp. By 2040, the global demand for new infrastructure investment is $68 trillion

Larry Fink annual letter

It's an immense opportunity that requires a new kind of investor, those that Brett Bivens and William Godfrey call “Production Capitalists”, firms that mix venture and credit and that can back companies in the long run through the different needs that growth brings.

Although it’s not going to be easy, and the challenge ahead of us is big and scary, we should not forget that energy drove what was the first attempt at a European common market, the European Coal and Steel Community - the EU as we know it now was born from the enlargement of the ECSC.

I consider the EU the biggest experiment of democracy, prosperity, and peace our world has seen so far, and it has its roots in energy.

As scary as it can be, I believe we are up to the challenge - let’s do it again, together.

Subscribe now

Read on and let me know what you think. Happy to discuss more at ricciuti_dot_fr_at_gmail_dot_com

My partner in ROI Chris edited this piece while wondering what's happening to journalism amid AI-motivated layoffs. Ping us/him if you have any thoughts...

Subscribe now

I usually resist generalizing anecdotal evidence to broad societal trends, but in this case it helps to convey my message so I’ll break my self-imposed rule. To my memory, no one in my high-school class, NO ONE, has built anything in the five years we studied together.

By building, I mean the old-fashioned practice of using your hands to work on or assemble a physical object.

If you’re wondering, I didn’t expect my friends to be modern takes on Leonardo Da Vinci and build a helicopter out of wood, but there are many things teenagers could dabble in like fixing a bike, creating small automations with Arduino, or simply building a treehouse.

And yes, I understand it might have to do with where we come from, with the kind of opportunities around us, and even with the physical space we grew up in, but it still feels odd to me.

Truth is that building stuff was not sexy anymore, and that makes a lot of sense given the service economy in which I grew up. Our parents were not building things, people we followed online made money with courses and drop-shipping, anything we owned was so cheap to buy that there was no point in fixing, and we therefore stopped thinking about building stuff and started thinking about building apps.

In an economy that outsources the building part and that is built around services, it's totally fine to abandon building and become a consultant, banker, accountant, marketing specialist, or maybe to learn how to code and become a software engineer.

Question from the public: “How far do you believe we are from having an AI operating 24/7 at the level of a junior engineer?”

Jeff Dean, Chief Scientist at Deep Mind: “Not that far. […] I will claim that’s probably possible in the next year-ish”

Source

Whether you think 12 months or so is accurate or not, the important thing here is not when it happens, but if it happens or not, and what happens next.

The way the software world is organized will change (or has it already?), and while I wouldn’t necessarily bet on how it will precisely change, I think it will directionally go towards a world with less “juniors”.

This fact probably has implications at every step of the ladder - every senior was once a junior after all - but we will start to see the effects on new grads first.

In fact, some say we have already started to see them:

And employers are saying the quiet part out loud: 70% say AI can already replace an intern. 78% plan to cut or eliminate graduate hiring altogether because of it.

Young People Can't Get Jobs. Now What?

Since the first industrial revolution, new technologies have gradually automated tasks previously done by humans - and have brought with them fears of unemployment. I’m a roboticist, and I’ve heard the argument that automation would be a net-bad for the society countless times.

Techno-optimists have always rejected these fears with the argument that technology doesn’t decrease the absolute number of jobs, but replaces automation-prone jobs with something more “human” and intellectually challenging. If this is true, they argue, automation just changes the nature of employment for the best, being a net-good for the society.

Although I have never seen data that undisputedly confirm or deny neither of the two theories, probably because of my cultural upbringing I tend to agree with the techno-optimists.

However, change of any sort - good or bad that is - brings with it opportunities. And this AI revolution is coming at a point in time in which the West needs change.

Our economies were built on hardware, but somehow we got lost along the way. Hardware is not sexy anymore - no wonder high school kids today prefer to spend time in the software world instead of building stuff.

How do we prepare them for a future where we need to get back to building?

One of the reasons why manufacturing tends to happen in clusters is that it is fueled by process knowledge, “the kind of knowledge that’s hard to write down as an instruction” (source).

It takes time to build process knowledge, and for it to persist an overlap between the old generation of workforce and a new one is required.

In the West, we have lost a lot of that process knowledge that we were so envious of, and that’s very clear in sectors that require a lot of it. Nuclear reactors, for example.

However, as Russian invasion of Ukraine has made very clear, bad things can serve as inflection points and necessity accelerates the formation of these process knowledge clusters: if we were talking about drones three years ago, nobody would’ve thought of Ukraine as one of the most advanced countries in the world at producing them. Yet here we are:

This is the opportunity I was talking about: getting juniors, those juniors that can’t find jobs anymore in the world of bits and software, to be excited about the world of atoms: this would help us reshore manufacturing, give fresh fuel to the diffusion of process knowledge, and would help us the sci-fi future the kids in the ‘60s were dreaming of.

How do we get kids to do something like this?

Before returning to my proposal, I want to address a common misconception.

By “getting back at building” I don’t mean this:

I mean this:

The way we build things has (luckily) changed, and now requires a highly skilled and educated workforce.

I’m not an expert in education, and I won’t venture in saying which is the best way to organize an education system to train these people. Germany has been successful with its Vocational Training System, while China has trained armies of engineers with very specific majors from university.

I’m fine with whatever works, as long as we get enough young kids excited to explore that route. You can have the world's best education system, but you will fail if your kids queue up the marketing department and don’t consider the rest.

This brings me to my main proposal:

A no-profit organization giving mini-grants to high school kids to build stuff.

It could work in the following way:

• €500 grants

• Kids would need to submit a detailed proposal of what they want to build and what they would need to do so

• No limits to what they can build but it must have hardware components, and the hardware needs to be built by the kids themselves

• As a consequence of the point above, we would fund only CAPEX

• Kids would agree to a deadline and a reporting schedule, as this would keep them accountable

• There should be a way to enforce the point above - not sure what that looks like, maybe dividing the grant in tranches and paying it out when certain milestones are achieved

• It would be great to provide them with mentorship of some sort. This would probably depend on how good we are in bringing people to contribute to the cause - with their time, if not with their money,

• It would be great to provide them with a physical space to use. First thing that comes to my mind is the worldwide network of fablabs - a partnership with local fablabs would be the best option to provide tools and knowledge,

• Since there should be some effort boots on the ground, and because of the point above, season 0 of the program would be focused on one city in one European country - unless it grows fast enough to allow boots on the ground in several cities at the same time

An initiative like this would help these kids get excited by the world of atoms, test their skills with things that most high schools in Europe (certainly in Italy) don’t train, and hopefully influence their choice of how to spend their time post-high school.

If such a thing takes off, there are thousands of different ways to make it evolve - I can think of partnerships with companies to provide tools (Arduino and/or Raspberry Pi: if you’re reading this, hit me up!), partnerships with companies to provide problems to solve, mentorship and internship opportunities, maybe even a small VC fund to back the craziest among these kids.

Or maybe there’s nothing behind it, nothing behind the main goal of having these young minds employed in the right sectors.

This is why the only metric I would track to define success is: how many of these kids, a few years after the grant, spend most of their time building in atoms instead of bits?

My pal Chris edited this piece while taking a break from his favourite new newsletter author, Anna Heim (TechCrunch), who's written some fun stuff for Resilience Media.

Everyone in VC agrees there's a big shift underway.

Whether we like it or not, we are seeing things that just a few years ago seemed impossible.

On one side, some companies became so big that they completely redefined the idea of what is a possible upside in venture. There are a handful of VCs still riding that wave, and raising huge funds.

A $20B fund sounds horrendously big, but when private companies can raise $40Bn at $300Bn valuation - as the mighty chat.com did - then returning such a fund doesn’t seem impossible anymore.

On the other side, there are people pointing out the arrogance of these funds. To deliver 2.5x on a $7B fund you need to capture around 30% of the total value created by the VC-backed startups during your investment period - which sounds unrealistic to be generous.

These people believe that there should be incentive alignment between GPs and LPs, and when you take 2% of $20B each year as a management fee, you have no interest in the carry anymore.

Truth to be told, many people smarter than me have noticed this shift long ago, before it became so evident to everyone.

However, I believe it’s worth discussing this topic once more because there's another shift happening, which requires a strong change in the mindset of venture investors.

If it’s true that big funds are getting bigger and bigger, it’s also true that it has never been easier to build a small, profitable and life-changing business.

First thing to understand here is that these gigafunds are playing a different game. It’s very easy to get hurt if you don’t understand the rules of the game you're playing.

Kyle Harrison has been active on this topic for years, so there's no need for me to elaborate much.

If Capital Agglomerators are targeting 2x funds, and Cottage Keepers have higher expectations for their small funds, at 5-10x returns, then we have a problem.

If both are competing for a different outcome, then one is willing to pay a higher price than the other. Here's the scenario: one Cottage Keeper and one Capital Agglomerator are competing to invest in a company with $5M ARR:

• Cottage Keeper: Bids $250M; believes the company can grow into a $2.5B outcome, and that could generate their 10x return.

• Capital Agglomerator: Bids $1B; believes the company can grow into a $2.5B outcome, and that could more than generate their 2x return.

Even if these two firms have the same fundamental belief in the company's outcome, their return threshold enables them to play different games.

The Puritans of Venture Capital

What this means is that the market is breaking in two, and as a manager of a small early-stage fund (the Cottage Keepers in Harrison’s piece) you have to recognize when a deal you’re considering belongs to the other half of the industry.

In some cases it is evident, but in others it’s not and sticking to your rules requires diligence and strength.

The main difference between the two worlds is that at early-stages gigafunds buy optionality, while small funds buy allocation.

For small funds it’s easy to fall into the trap of high valuation, competitive deals, thinking that even if you don’t get to your target allocation now you can still invest more later on. This is a big fallacy: if a deal is worth investing in in the next round, it will be too big for your small check to make a difference, while if your small check can make a difference in the next round then it means it’s not worth investing more.

At this point, a clarification is needed.

I know a few fund managers who reading these words would nod with satisfaction, thinking that they understand this and that’s why they are building a portfolio of safe bets: “I won’t invest in the next OpenAI, but that’s because I’m playing a different game.”

I’m sorry, but that’s bullshit.

There’s never been anything such as a “safe bet” in venture capital. The power law is the single reason VC exists as an asset class. Most of us won’t even return the money LPs invested but, the few that will, will make their LP very, very happy.

AI is making the concept of “safe bet” even more stupid.

Since it has never been easier to build a small, profitable and life-changing business, even if a low-hanging fruit existed, it would be picked by someone building a lean, AI-powered, probably profitable company.

Sometimes we get trapped in this world of billions and trillions, and tend to forget that selling a business for €5M is way easier than selling one for €5B, and €5M is enough for most of us who strive to have a peaceful and satisfying life.

So now, in an unstable, multipolar world where will bright, ambitious people turn next? Now that Facebook and Google are uncool, interest rates are high, software competition is going infinite, and institutions feel unreliable, what is the next high status path that will dictate our economy?

They’re going all in on cashflow and independence because they’ve learned that institutions won’t take care of them and can’t monopolize opportunity.

Free cash flow and independence (two sides of the same coin) are hedges against uncertainty, institutional decrepitude, and burnout.

Yoni Rechtman

All of a sudden those fund managers that believed in safe bets are seeing their spot in the world shrinking, eaten from above by gigafunds buying optionality and from below by ambitious people who JDGAF about these fund’s “value add” and are building on their own.

What’s left?

Simple. The crazy bets. Those deals that are so ambitious that seem impossible, so far-fetched that only venture capital, in the literal meaning of the word, can finance them.

If you are not brave enough to invest in these kinds of businesses, there is no space for you in this new era of VC.

Now more than ever, you can either go big, or go home.

My buddy Chris Spillane edited this edition of Unbeaten Path amid his shameless gloating on social media about this newsletter's growing popularity.

Thank you for your support so far!

VC is an industry that lives off bubbles and hype cycles - there's nothing that can be done about it because it’s baked into the model.

The latest hype is about the physical world, manufacturing and more generally, the re-industrialization of the West. Although I am sure many are driven by genuine interest in the topic or high ideals like patriotism and security, I question how they come to articulate their investment thesis.

Much of what I have read around the web is full of optimism and grandiose statements, but fails to grasp the reality of what manufacturing really means, what re-industrialization really implies and, most importantly, the consequences for the VC industry.

I'm no manufacturing expert - in fact, I have never worked in a factory, nor have I been responsible for the high-volume production of any physical goods. For this reason, I will try to articulate what the manufacturing opportunity is using common sense and real-world data as much as possible. The hope is to scratch a bit deeper than what hype-filled LinkedIn posts usually do.

The first note I want to make is on the very definition of what “re-industrializing” the West should mean and why tariffs à la Trump are stupid.

A perfect trade balance of goods is not what we want, as it creates weird incentives – $1M worth of socks is not the same as $1M worth of semiconductor equipment. Certain things are strategic for a country and pleasant to work on; some others are not, and the West should aim to re-shore the strategic and pleasant part of manufacturing that attracts and retains talent and nurtures innovation.

By pushing for a perfect trade balance of goods, you push other countries to treat socks like manufacturing equipment - and that is not good if you don’t want to manufacture socks.

And that’s the frustrating part - we shouldn’t dream of making socks here. The real pivot to the US is advanced technology like AI, chips, biotech. By decoupling or punishing foreign partners, we might accelerate them building their own R&D ecosystems. If they succeed, they’ll rely less on this really important American technology. Then we’re not just ceding a shoe factory or whatever, we’re ceding the next wave of innovation. That’s a deeper, darker implication.

Bonds, Stock Market, and China: What's Going On

What seems like a superficial consideration actually has some deeper implications.

As I wrote in “Europe, it’s time to build”, the main issue with China is that it has moved up the value chain and, by controlling the foundational layers of it, now is able to own also the highest value-added piece of the chain – the strategic and pleasant part.

Western companies’ strategy to delocalize the low value-added pieces of the chain has worked amazingly for long - and we shouldn’t be surprised it did!

By making the hardware on which their SW runs cheaper, these companies have commoditized their complements. If you’re Apple, you make a 50% margin on your iPhones, but you also make much, much more from the software ecosystem you have built around it. In this case, the iPhone is a complement of the App Store, cloud services and the rest: since the demand for a product increases when the price of its complement decreases, if Apple can sell its low-margin hardware products for less, the rest of their offering benefits from it.

This strategy works damn well - no wonder Apple had a market cap of $360B when Tim Cook became CEO (and started Apple’s Chinese strategy) and has reached a peak of $3.5T in 2024.

Now, however, Apple is losing market share to Chinese competitors in many markets - the same competitors that were once suppliers and that have moved up the value chain. Maybe now it would make sense for Apple to reshore part of its production to protect the advantage it still has and to keep the products of their USD-fuelled R&D away from its enemies’ eyes, but it can’t because the US has lost know-how and wouldn’t be able to produce the same things China does.

This consideration brings us to the core of what re-industrialization is all about.

Service economies are different than manufacturing economies, and shifting from a service-heavy economy to a manufacturing-heavy economy requires changes at many different levels.

I believe this shift is an inflection point, an exogenous shock to a well-oiled system that opens up opportunities both in markets that are directly affected and in many other adjacent ones.

The most obvious beneficiaries of this shift will be the companies directly tackling the re-industrialization needs, which I will call “the primes” now on.

The primes take the spotlight, tackling the biggest opportunities and making great stories.

Primes in robotics and automation, energy, defense and computing have a great opportunity in front of them: disrupting large markets dominated by slow incumbents by leveraging this inflection point.

Often these companies look very similar: they are vertically integrated, they make extensive use of technology to build their product, they are lean and move fast, first tackling these markets from the edges and later expanding their scope, eating market share from incumbents.

Moreover, they have the benefit of having to set up their supply chain from scratch and thus don’t have to go through the painful process of reshoring or rethinking their relationships with suppliers.

The opportunity is so big that this is what most of the funds targeting re-industrialization are financing. These companies might be capital intensive, but if financed early enough, they can be exceptional investments.

So much has been written about why vertical integrators win during inflection points (including something by me) that I won’t add more on the topic here.

However, for these primes to be successful, there must be a broader change in our economies, providing the support they need to scale.

I, Pencil, simple though I appear to be, merit your wonder and awe, a claim I shall attempt to prove. […] I have a profound lesson to teach. And I can teach this lesson better than can an automobile or an airplane or a mechanical dishwasher because—well, because I am seemingly so simple.

Simple? Yet, not a single person on the face of this earth knows how to make me. This sounds fantastic, doesn’t it? Especially when it is realized that there are about one and one-half billion of my kind produced in the U.S.A. each year.

”I, Pencil” - Leonard E. Read

The main question is: how vertically can the vertical integrators integrate? How much can they build by themselves?

The complex thing about supply chains is that they are… complex. Every component of a product consists of other components, each of which has its own supply chain. If you go down to the raw material, you will face the same problem: where do we source the machines to extract the raw material, and where do we source the pieces to build the machines to refine the raw materials?

Take Tesla, the poster child of vertical integrators.

One of their innovations that fascinates me the most is their Gigapress, a giant robot shaping aluminum sheets in one go. Well, Tesla doesn’t build that robot on its own: some of them are built by an Italian company, Idra Group, and others by a Chinese company, LK Machinery.

Even the most die-hard vertical integrators will have to source machinery, materials, and energy from somewhere. That’s where a service economy will fail to provide what they need.

Robots are a good example to show the problem.

China is leading the way in robotics, and we are lagging behind.

China invested heavily and now is starting to export its robots here.

Coco Robotics, which has just signed a global partnership with DoorDash, built its fleet of robots on top of a Chinese-made chassis. Chinese producer DJI is the world’s biggest manufacturer of drones.

This opens up a big opportunity for robotics primes, which often vertically integrate, going end-to-end and solving a specific problem – take Monumental, Exotec, or Verity as examples.

These primes build their own robotics platforms; they own the software stack and take care of deployment themselves.

However, we shouldn’t forget that building a robotics company means sourcing raw materials, batteries, sensors, and actuators.

Unfortunately, we can’t produce these things on European or American soil.

In 2024, for example, China produced 44,700,000 tonnes of aluminum, while the USA produced only 1,360,000 tonnes. What are we going to build robot frames with?

Another stunning example is magnets for electric actuators, which are key for automotive, robotics, and defense but also for semiconductor manufacturing. Well, China has a complete monopoly in heavy rare earth metals and produces 90% of the world’s rare earth metals - materials required to manufacture the magnets in electric motors.

What happens when China cuts the supply?

China has suspended exports of a wide range of critical minerals and magnets, threatening to choke off supplies of components central to automakers, aerospace manufacturers, semiconductor companies and military contractors around the world.

Shipments of the magnets, essential for assembling everything from cars and drones to robots and missiles, have been halted at many Chinese ports while the Chinese government drafts a new regulatory system. Once in place, the new system could permanently prevent supplies from reaching certain companies, including American military contractors.

https://www.nytimes.com/2025/04/13/business/china-rare-earths-exports.html

Other examples of areas where China leads include sensors like cameras and lidars, as well as batteries that power these robots.

Even companies operating in the most sensible sectors - such as defense in the US - can’t build without sourcing components from China.

If we really want to go down the route of re-industrialization, then all these sectors become “investable” sectors with big growth perspectives.

Continuing our example on the tech stack needed to build robots, we find chips giving them intelligence.

Also here, the primes are the ones taking the scene - have you ever heard of this tiny company called Nvidia? But there are monopolies at every step of the chain.

This is probably one of the most fascinating industries to study. People like Chris Miller and his book “Chip War” have helped educate the public on the importance of certain single points of failure.

To make a chip, it takes many steps, from sourcing and refining silicon into 99.9999999% purity to printing, testing and packaging, and no country can really make it on its own.

It’s important that we understand how our countries need to change to be more resilient and sustain the re-industrialization efforts properly - what happens if China invades Taiwan, or if the US decides they won’t export quartz from Spruce Pine anymore?

Each one of these single points of failure represents an opportunity for those who are brave enough to invest.

Energy is another obvious example. Manufacturing is power intensive, and a manufacturing-heavy economy must be able to provide its industries with cheap and reliable energy if it wants to be successful.

Also in energy, there are some obvious needs that are being addressed by the primes, like energy production or energy storage, for example. But there are also some second-order needs like reliable grids and more efficient energy markets, that still make big and interesting opportunities, and on whose growth depends the fate of the primes.

Energy, like automation, is a sector that is highly dependent on a few technologies that the West doesn’t control. Let’s take two examples (but we could make many more): transformers and photovoltaic cells.

Transformers are boring but essential components of our grids. Won’t digress much on what they do, but if you feel a little nerdy, check here for a good article. A shortage of transformers has been slowing down the development of new energy projects and, in general, any project that needs lots of energy (think datacenters, for example).

The problem is that Europe and the US can’t produce enough of them because we lack skilled labor and we lack raw materials needed to build them - China alone produces 50% of the world’s electrical-grade steel.

How do we build offshore wind farms, solar farms, nuclear reactors, datacenters, and anything that needs energy if we can’t build enough transformers?

Solve this problem, and you’ll have a very valuable company.

Photovoltaic (PV) cells are another unsung hero of our energy system, but also another piece of the stack that the West doesn’t control.

China has nearly a monopoly in this sector because producing photovoltaic-grade silicon is very energy intensive, and they have built enormous production facilities near huge power plants that supply them with cheap (often renewable) energy.

Energy-intensive processes are, by the way, also one of the reasons why China is leading in the production of many other raw materials such as steel and aluminum.

China is, for example, the biggest exporter of iron and steel articles – structures, pipes, fasteners, springs, etc. – with around 25% of the world's total world’s export amount.

The interesting thing is that China doesn’t even have enough domestic iron ore production to fulfill its needs and imports around $112B of it a year, mainly from Australia.

They managed to become top exporters of products made of a material they needed to import from other countries because they mastered the refinement process (with a little help from the government).

Steel is absolutely essential if we want to get back at building:

Not everything in the world is made of steel, but nearly everything in the world is made with machines made of steel

Ed Conway, “Material World”

How can we catch up?

Notice that most of the chemical processes to manufacture these materials were invented in the late 1800s or early 1900s: Bessemer process – 1856, Siemens process – 1953, Czochralski process – 1915, etc.

Is there a better way to produce these materials? Can we find more efficient steps that don’t require high temperatures? Can we invent better catalysts to facilitate the chemical reactions?

We must, if we want to fulfill the energy needs of a manufacturing-heavy economy.

This shift to manufacturing-heavy economies also implies shifts that are not directly “VC opportunities” but that could be growing markets for many companies to find customers with a high willingness to pay.

Think about infrastructure, for example.

With all of the talk of reshoring American industry and building manufacturing capability, one target of this effort should be on public infrastructure. At times overlooked for its more visually-appealing cousin, manufacturing, infrastructure builds the foundation with which manufacturing can take place. I’m talking about roads, bridges, electric transmission, water desalination & piping, dams, deepwater ports, airports, high speed rail, and more.

Someone Please Build This: America’s Industrial Construction Co.

An increased need to move things around is a second-order effect of an industrial economy: you move energy, you move people, you move goods, and you move heavy machinery. To move things, you need infrastructure.

To see what the creation of a manufacturing economy looks like from an infrastructure perspective, check these graphs from … you guessed it, China.

It's no secret that China is years - perhaps decades - ahead of the US in building the aforementioned infrastructure projects, and the US needs to catch up. While I won’t go into all of the reasons why, the fundamental argument here is that the reshoring of American industry cannot happen without infrastructure to support it. The most efficient, automated factory in the world is hamstrung without a proper road or rail network for transporting raw materials in and finished goods out. The most awesome nuclear plant, built at record speed and cost, has no way to get its electricity to millions of homes and businesses if there isn’t a highly capable transmission network. I could go on…

Someone Please Build This: America’s Industrial Construction Co.

We need ports and trucks and rails and ships.

Are we ready to get back to building them?

Subscribe now

“And just like cigarettes, we won’t work with anything that’s… evil.”

Evil is a powerful word. It’s usually reserved for the worst of us, or the “profoundly immoral and wicked”, according to the dictionary.

Nonetheless, those were the words of a key decision-maker at one of the world’s largest tech conferences. And they were in response to my question on managing the increasingly blurred lines between deeptech, dual-use, and defense tech.

I have a strange relationship with defense. I guess that’s what happens when you’re thrown into a North London karate dojo at age 5. I’ve always felt the right to defend yourself is absolute, so much so that I only stopped boxing a few years ago.

Most people view defense as a euphemism for the military. Military is aggression. Aggression is violence. Violence is wrong. Wrong is evil. Frankly, military action in my lifetime has done very little to arrest this linear thinking, so it’s no surprise that it doesn’t take much for the word “defense” to appear next to “evil”.

But like people’s minds, the world changes. However, prevailing attitudes towards defense seemingly haven’t, and people, particularly Europeans, are still squeamish about it. It’s a perception that Europe’s reawakening defense industry needs to address with immense urgency.

I kicked off with an anecdote, so here’s another one. I was recently in London at the Future of Computing conference and mentioned in passing (over a pint, of course) that many of these next-gen computing platforms will have defense uses. A person I didn’t know interrupted me, saying, “It’s disgusting that national health spending and pensioners money are going towards missiles.”

To be fair, pensioners’ money has historically funded worse ideas, like Eurovision. Anyway, defense is not (just yet, at least) being subsidised by pensions, but why is it tough to change the prevailing sentiment? People are nervous about what defense and adjacent categories truly mean. How to fix it? Maybe the answer is hidden in the following.

Art Imitates Life

Every Mafia hit job that lives in infamy in Wikipedia or documents is accompanied by at least an hour of exhilarating, expertly crafted storytelling in The Sopranos. The Vietnam War inspired some of the greatest films of all time. Political dramas like 24, Homeland, or even The Bodyguard followed the 9/11 terror attacks, rarely deviating from terrorism-related storylines.

What was Tom Cruise before Top Gun? Would Tom Cruise be who he is today without aerial sparring partner Val Kilmer? (and I write this as someone who’s not a Top Gun fanboy... There’s no need for a gritty reboot where The Cruiser pilots drones from his Berlin co-work while sipping a cortado and complaining about GDPR.

Each of these shows has helped rebrand their specific thematic area in a more positive or comfortable light. And Europe’s defense tech sector may need to do the same. Rebrand from evil to essential

The defense industry urgently needs a rebranding from a perception of aggression towards one emphasising protection, stability, and human dignity. Technology built to defend can just as easily protect civilian lives, secure vital infrastructure, and support humanitarian missions.

This started quietly in 2022, with the European Defence Fund committing nearly €8 billion to collaborative defense research from 2021 to 2027. The Fund focuses on projects that provide tangible civilian benefits, like improved cybersecurity, secure communications, and disaster response capabilities, effectively reframing defense from aggression to resilience.

For instance, during recent flooding events in Germany and Belgium, defense infrastructure and military personnel were instrumental in crisis response and civilian rescue operations, illustrating the essential humanitarian role defense technology can play.

But like many European Union-led efforts, the storytelling and PR chutzpah were abysmal. They were way too quiet. So when European Commission President Ursula Von der Leyen outlined her ReArm Europe vision with headroom to spend €650 billion, no wonder people (read: Europeans) were shocked.

For context, that’s approximately four PSG squads or half a flat in central London.

However, founders and investors must openly champion these narratives, emphasising how defense investments directly contribute to stability and protection rather than conflict, all for the benefit of the economy.

The Duality of Defense Tech: The Knife’s Edge

Defense technology inherently possesses dual-use potential. It’s capable of both profound protection and harm, dependent entirely on intent. Technologies, like drones, artificial intelligence, and photonics, epitomise this duality.

European startups increasingly recognise the importance of explicitly defining ethical boundaries. Take Helsing AI, for example. As one of Europe’s most impressive defense tech startups, it exemplifies this by emphasising their mission "AI to protect our democracies" and explicitly ruling out surveillance capabilities like facial recognition. Instead, they focus strictly on defensive applications, such as detecting hostile drones or threats to civilian infrastructure. Furthermore, the German startup will only sell to democracies.

European policymakers reinforce this duality through stringent regulations and ethical frameworks. NATO’s principles of responsible AI use, mandating human oversight and accountability, further underscore Europe's commitment to ethical boundaries. For investors and founders alike, clear ethical delineation is crucial to public acceptance and market viability.

Defense Tech and Public Trust: Fear Won’t Fund the Future

Public trust is vital for European defense startups to thrive, and transparency is the pathway to gaining that trust. Historically, ambiguity surrounding defense technologies has fostered suspicion, which complicates investment discussions and public acceptance. Successful European defense startups understand the power of clear, human-centric storytelling.

For example, Dedrone, a German company providing drone detection technology, consistently highlights their commitment to protecting airports, stadiums, and civilian areas from rogue drones, thus reframing their product as a vital public safety tool.

Such approaches demonstrate the effectiveness of transparency in reducing skepticism. By proactively communicating their missions, values, and impacts, defense startups can dissolve fear-driven narratives and open the door for wider societal support and investment enthusiasm. Clear, authentic communication not only helps explain technology; it also humanises and contextualises it.

The European Paradox: Defense Shyness in the Shadow of War

Europeans remain hesitant to fully embrace defense investment, despite the stark geopolitical realities highlighted by events like Russia’s invasion of Ukraine. This hesitance arises partly from Europe's unique political structure.

The EU isn't a nation-state and thus lacks the unified nationalism that often underpins defense narratives in countries like the United States. American defense investment is typically rooted in patriotic sentiment, while European initiatives must appeal to broader values of collective security and shared democratic ideals.

Ironically, even historically non-nationalist Canada recently experienced a surge of patriotism around defense and sovereignty, leveraged effectively by Mark Carney to clinch an election victory, suggesting defense needs can spark nationalism where none previously existed.

It appears Trump doesn't just win elections in the U.S.; his influence on global defense discussions indirectly secures victories elsewhere.

For Europe, navigating this paradox requires reframing defense-as-a-collective (DaaC, anyone?) and practical responsibility rather than an ideological battleground. Emphasising strategic autonomy, safeguarding democracy, and preserving peace offers Europeans a narrative they can collectively rally behind without succumbing to traditional nationalism.

Defense isn’t inherently evil, it’s urgently misunderstood. Europe’s future hinges on reframing defense as a shield rather than a sword. If European startups and policymakers master storytelling, they can demonstrate how defense tech protects the very fabric of democratic society, from safeguarding critical infrastructure to ensuring humanitarian aid reaches those in need.

Europe’s squeamishness about defense is a luxury it can no longer afford. As geopolitical fault lines deepen, failing to clearly differentiate defense from aggression is strategic negligence. Founders, investors, and policymakers must urgently rewrite the script, framing defense as a fundamental necessity for peace, resilience, and sovereignty.

The stakes are too high for quiet diplomacy or PR modesty. It’s time that Europe embraced the uncomfortable truth: sometimes, the greatest evil is doing nothing at all.

My buddy Chris edited this edition of Unbeatean Path in Vilnius while listening to software engineers explaining how they are building the next-generation of defense drones

Venture capital is a long game, but it's also highly opaque, and the two are actually connected.

Being a long game, it takes time between investments and cold money in the pockets. During this time, venture often resembles a Ponzi scheme: as long as someone invests at a higher valuation, you mark up your position and pat yourself on the back while admiring your TVPI. You literally have no incentive to mark that position down, unless there is some hard signal that you must do it - a.k.a. down round.

As a result, it takes a long time before a bad fund is actually recognized as a bad fund.

Startups raise money every 18 months, and every 18 months they must face the market and its judgement: sometimes they manage to increase their valuation even without proper traction but, more often, their valuation will take a hit (or they won’t raise) without results to show. VCs, on the other hand, raise funds every 4 years and often have markups on paper to show as a promise of future performances. This means it will take at least a couple of cycles of fundraising for the fund to change its reputation.

Just to be clear, this is not a criticism of the system – it’s the way VC works – and good portfolios take time to mature before they show results. However, an undeniable consequence is that the “quality” of a fund is not simple to assess before almost a decade has passed.

Moreover, even if startups or other fund managers wanted to assess the quality of a fund that's old enough to have a mature portfolio, it would be very hard, as their performances are not public (except for rare examples like the great team at Unruly).

The closest proxy for performance we have is their portfolio. One could theoretically look at a fund’s portfolio, look at when they invested in each company, make an assumption about ticket size and valuation, and make an educated guess.

This is certainly better than nothing, but anyone who has spent some time in venture knows that logos on websites never tell the whole story, and there are just too many variables that could change the estimation substantially.

Take this example:

The firm says it has achieved 17 exits so far, the largest of which came from speedy grocery giant Getir. Revo first backed the company in 2017 and then cashed out in a secondary sale at the height of the speedy grocery boom in 2021, when Getir was valued at $7.5bn. The firm did not disclose the size of the stake it sold but Sifted understands it comfortably returned the value of the firm’s first fund. The VC offloaded its Getir stake at an opportune time. By 2023, Getir’s valuation had fallen to $2.5bn, and last year it withdrew from all markets outside of Turkey.

https://sifted.eu/articles/revo-capital-closes-86m-fund

If you are the Turkish VC Revo, you are very proud to have Getir on your website. However, you have to specify that you sold when the company was valued at $7.5bn, because now it is valued at $2.5bn - one-third of that. But what if you are Tiger, who invested in the startup during their Series C at $2.6bn?

Moreover, as Sifted notes, the size of the stake matters.

According to Dealroom, Revo invested in Getir’s $38M Series A. Take the hypothetical case in which the Series A valuation was $150M and imagine the round was oversubscribed and there was only $500k of allocation left: that investment of $500k at $150M would have been worth $25M at the home run valuation of $7.5bn, which is unfortunately only a quarter of Revo’s $100M fund.

On the same theme, there are several examples of sub-$1bn exits that return funds multiple times because of the entry valuation, cash efficiency of the business, and fund size.

The main question I am trying to answer with this long preamble is: what makes a VC fund “Tier 1”?

If you can’t assess the performances of funds based on real data, it takes time for portfolios to reach maturity, and the conditions when one deal was made are so important - how can certain funds be widely recognized as better than others? why is it that some of these funds give strong signals to the market whenever they invest? why are some funds “Tier 1”?

Frankly, I think venture is a brand game

Taavet Hinrikus from Plural at 20VC

And, frankly, I agree.

My thesis is that in most cases, the brand comes first and the fund’s performances follow - not the other way around. And there is nothing wrong with it.

If you are a returning reader of my work, you know how much emphasis I put on the importance of storytelling. I believe that storytelling is a self-fulfilling prophecy, and, as such, building a strong and credible brand around yourself is enough to convince people to think you know what you are doing. Once people think you know what you are doing, they follow your guidance; the best investors want to invest with you, and the best founders want to work with you.

If you manage to put yourself into this positive cycle, you carve your spot in the industry, and the rest comes.

Don't get me wrong; there is much more to venture than sourcing - after sourcing you have to select the best deals, secure your spot, and support the founders as much as you can.

But I think the 4-S of venture follow each other in a precise order, and as in sales you will fail if the top of your funnel is low quality, also in venture it all starts from seeing the best startups first. Without seeing the best, you could be amazing at selecting, but you would still fail – the power law is reckless, and selecting the best among mediocrity doesn't move the needle.

Can someone have the “magic touch” for sourcing? A secret sauce to succeed?

Maybe back when Warren Buffett was counting tank cars in rail yards, yes. But in the era of data-driven strategies, AI has leveled the playing field, and secret sauces don't exist.

It's not by chance that most of the emerging managers these days are very similar on this front: all of them have spent time building an audience, all of them share their thoughts in long or short form, and use their brand as a competitive edge.

In most cases, the audience comes first, and investing is almost a byproduct - a way to monetize the work done to build a following.

I think VC has always been a brand game - both Paul Graham and Fred Wilson started writing in 2003 - but for a long time there was another way of doing things, a more traditional way.

I don’t think the traditional way works anymore.

Subscribe now

My buddy Chris edited today's Unbeaten Path from the Catalan capital, Barcelona, where it's Sant Jordi - a day where locals give their loved ones a rose or a book.

What are the chances all these people chose the same subject line for their cold outreach? AI outreach is clearly out of control here.

Leaving my irritated-VC comments aside, this made me think: what’s more important, the magnitude of your intelligence, or the delta between your intelligence and your competitor’s?

This is somewhat the same question that is at the core of the controversial Enhanced Games - privately-owned event that's inspired by the Olympic games but doping is accepted: if each athlete is doping, then it’s like if none of them took drugs. In other words, performance enhancing drugs increases the magnitude of the performance, but the delta doesn’t change.

To rephrase this in more generic terms: in a world in which intelligence is commodity - there are almost one million LLMs on Hugging Face - how can it be a differentiating factor by itself?

Sure, there will be some transition period in which the magnitude of intelligence will make a difference, and those that use AI will outpace or outwork those that don’t. But at steady state, it's the delta that's the difference maker, not the magnitude: while the magnitude of intelligence is artificially increased by AI, it's how well you apply that intelligence on the ground, the little things you know that others don’t, that makes a difference.

I believe this applies both to company building and to venture investing.

Back when building a data-driven strategy was a job for specialists, having a smart scraper that could get the right signals from the web might have been a competitive advantage. Today, at least from my experience, it won’t make a difference.

Anecdotally, a few of my portfolio founders were highlighted by data-driven VC platforms and were flooded by messages. They prioritized replies exactly how you would expect: brand and connections.

This brings me to an important point: in a world where intelligence is a commodity, it's brand and connections that will make a difference between out-of-distribution and average performance in venture investing.

Brand has always been important in venture. There is this feeling in this industry that certain people/funds are the hidden hand, a kind of King Midas of VC, and whatever they associate their brand with becomes the next big thing.

The best way to escape this has been for a long time to invest in small pockets of the market where deals are illegible to most - either because they emerge from a little understood geography, the topic is highly technical, or the stage is extremely early (oftentimes before the company even exists). If you did this well, you could pick a few winners and rise to become the kingmaker. This is, for example, the story of Hummingbird - “The best venture firm you’ve never heard of” - that is now a household name and its presence in a deals signals quality and exclusivity.

With intelligence increasing in magnitude, however, it is sufficient to change your LinkedIn bio with “Stealth” to be contacted by dozens of investors. There still are very illegible deals, but have become very rare and it’s probably hard to find pockets of the market that are still undiscovered.

This is where branding and storytelling come into play.

Are you top of mind for founders and other investors for what you do? Do people know what you think, and is it easy for them to identify the deals you are interested in? What are you doing to build such a brand?

These are all questions investors should ask themselves, as those that don’t will end up underperforming. This is also one of the reasons I write here: despite having a small audience, every piece I write gets me connections with like-minded people, and helps me build a brand that people recognize.

The need for connections is also nothing new, but the way networking works is changing fast.

As AI eventually kills outbound as a channel, networks will become all-important.

The scale of outbound sales will increase by orders of magnitude and the well will become permanently poisoned. Your bot’s emails and calls will just go unanswered. Your AI generated SEO blog posts will go unread. Your personalized video won’t feel special.

There will be infinite, undifferentiated competition for every opportunity and so those direct with access to wide or specialized networks of buyers, talent, knowledge, etc. will become valuable gatekeepers. Distribution will be everything.

Yoni Rechtman - Get Ready to Pay Your Network

Sometimes this job feels very transactional, and are those people that are able to build a real personal connection that will stick in my mind. Those people do little things in an exceptional way - they remember your kid's name, for example, or they recall you were sick last time and check in on you.

Creating strong personal connections is a superpower, that in an age of AI-powered noise becomes a differentiating factor. Digital social networks are dying under the choking weight of AI, and physical social network - the real human connections - are going to be a competitive advantages.

A second-order effect of physical social networks is that they are constrained in size by definition, as it’s impossible to keep in touch and build real connections with too many people. This makes specialization even more important, especially at the early-stages. If you want to build a network of people that deeply understand the space, that provides a competitive edge, you have to use your time wisely and spend it in the right places. Do you invest in highly technical topics? Then it’s probably better you spend your time at academic conferences rather than giga-events that favor quantity over quality.

This is a great opportunity for those early-stage investors that are willing to put in the work, willing to specialize and take risks others are not willing to take. Maybe this is the reason why there are many first-time managers targeting very specific thesis and not going wide but going deep.

I think this is the right way to go.

My amigo Chris edited this version of Unbeaten Path while trying to decipher tariffs with The David McWilliams Podcast. It’s less to do with economics and more to do with culture wars.

Finally, let me say this: No CEO with an ounce of intelligence will invest capital to buy thousands of humanoid robots to work in their factories or warehouses. They will redesign the process instead to eliminate the use of humans altogether.

Brittain Ladd

Brittain Ladd is very opinionated with the added bonus that’s he’s often right. The comment above is what many I know in the robotics industry think but maybe lack the bravery to share with such a big audience.

Why? Well, what if they’re wrong? Being wrong is bad and doing so in public is horrible. Sadly for venture investors their opinions are expressed via their investments, no matter how discreet they are. There’s no way out of it, so this is what I think of humanoid robots.

They don’t make much sense.

But then, why is the industry - or at least the loudest voices - focusing so much on humanoids?

Jordan Nel nailed the reason in his piece “Robotics: A Product Selection Problem”: the TAM for humanoids is huge, literally every human activity.

Today they are starting with boring warehouse automation tasks because they can operate in a safe environment, without humans, on tasks that are not safety-critical and can provide valuable feedback to engineers. The real goal, however, is to bring these robots in the real world, replacing waiters, nannies and even postal workers.

I don’t buy this vision of the future.

Those who subscribe to this vision are committing the same error as the founders of Humane. They believed that what their small bubble of rich, tech enthusiast, Silicon Valley-based, and Twitter-educated people thought applied to broader humanity. We should never forget that iRobot’s Roomba, a very basic robot doing only one thing, that can be bought online for a few hundred bucks, that can save us from the hated task of cleaning the floor, sold “only” 50 million units since 2002 and the company was valued at $1B during Amazon’s failed acquisition. This isn’t bad in absolute terms, but it’s tiny for a consumer product, a leader in its space and sole player for a long time. For comparison, Apple sold 2.3 billion iPhones since its launch in 2007.

The belief in such an immense TAM is however very appealing for a certain type of fund, those that have billions of dollars to invest and need such humongous outcomes to justify their existence.

The reason why I am bothered is because I think we are missing a big opportunity here.

We are in a special moment in history. A period of change, a time of disruption. Industrial tech is becoming strategic and robotics can play an important role in making the West competitive again with China (as shared here, and here, and here).

Increased intelligence is one of the inflection points that will allow startups to disrupt large markets dominated by slow incumbents.

If we dedicate too much attention and money to the wrong applications of robotics, we could waste the momentum and miss the opportunity to use them for what they do best: dirty; dull; and dangerous jobs.

Ladd’s earlier comment is from a post where he opined about BYD’s new gigafactory, which is supposedly bigger than the whole of San Francisco.

BYD and many other Chinese manufacturers build such huge factories because there is value in vertical integration. These factories are highly automated and employ many robots. In some sense, they all learned from Tesla. When watching the infamous fly-through of Tesla’s gigafactory the thing that strikes me the most is how little welding there is. One of the most valuable innovations that Tesla brought to the car manufacturing world is the Giga Press, a gigantic, single-purpose robot that shapes metal sheets in one go. After Tesla introduced this technology in their factories, Chinese manufacturers quickly did the same and now also traditional carmakers are adopting the tech.

Do you see the parallels here?

For decades, we have produced little pieces in separate processes (and often in different parts of the world), welding and assembling them together in a long and inefficient process. We could’ve replaced humans with humanoids in those same production lines, but that would have been completely stupid as we would have been optimizing for a local minima - i.e. the human-centered factory - instead of escaping that local minima redesigning the whole process in a better way.

The key to making robotics work is starting from the problem, not the process. Processes are just solutions to problems. Processes can be changed if there is a better solution to the problem.

A typical (and overused) example here is the car vs horse: the problem is to move people faster and more comfortably, and the solution is changing the mean of transportation and not making horses faster.

To generalize, there are many problems that can be efficiently solved by specialized, product-centric robotic solutions. If a solution is offered at the right price, it solves a problem that’s big enough. If it actually works, customers are very happy to change their processes to adopt it.

Solutions like Autostore, Exotec or Kiva, for example, changed the way warehousing looked like. They did it by thinking about fundamental problems first, and then about the technical solution to optimize that process.

In other words, why should I use the same humanoid robot to assemble a car’s engine, prepare a poke bowl and sort goods in a warehouse? When all you have is a hammer, everything looks like a nail. If you massage the topic enough you can find similarities between the above-mentioned processes. However, the real world is different, these processes are very different and the right approach, in my opinion, is to have vertical solutions building processes tailored to the problem at hand.

Isn’t this the way it has worked for a long time? Yes, it is - but there is a caveat. Today intelligence is cheaper, more flexible, and easier to deploy.

I started working on autonomous driving in 2017, and I vividly remember when Wayve made waves (no pun intended) in the industry by claiming that they could do autonomous driving with end-to-end AI. People thought it was impossible, that it didn’t make sense, that “those who knew how to do it” were doing it in another way. History tells us Wayve was right: now everybody - Waymo and Tesla in primis - adopted the same approach.

Why did Wayve’s approach win? Because it was cheaper, more flexible and easier to deploy. All of a sudden, you can train your car in London and drive the next day in Berlin and Los Angeles.

The same thing, I believe, is happening to other areas of robotics. Today, robots are able to understand the real world way better than they could do before, and all of a sudden it’s not unreasonable to think about robotics deployed by traditional SME, increasing their margins and optimizing their processes. The vertical robotics solution - purpose-built with specific problems in mind - can now give superpowers to traditional industries, which are those most affected by aging population and lack of people willing to do the job.

One word of caution is needed though.

After Nvidia GTC, the VC world has been flooded by the concept of “Physical AI” and many software investors that missed the “generative AI” train boarded this one as if it was the second coming of Christ. The problem is that the real world is not software, and “good enough” in the real world doesn’t have the same meaning as good enough in the software world.

At my first startup, we were building autonomous robots and we had a robot that was running around our office autonomously every day, the whole day. Once in a while something minor would happen, but we were happy anyway: it was good enough for us.

Somehow we managed to convince a very very big US retailer to test our robots, and soon we understood what “good enough” means in the real world: >99.9% uptime. Needless to say, that wasn’t even close to what we had, I wouldn’t be here writing this post otherwise.

This little tale taught me something: in robotics, your competition is usually physical labor, which generally works well enough. The question “Does it work?”, in robotics, has a whole other meaning than in software.

“Does it work?” is the first question one should ask when doing due diligence on these companies. It shouldn’t be answered superficially.

If you’re building a full stack company, your TAM, value proposition, and customer “orders” will typically be obvious. But, we care about performance. That’s the hard part. And that’s why the opportunity for full stack remains massive — and underpenetrated.

Does it Work? - Eclipse

Investing in the real world is a huge opportunity, but building in it is very delicate.

As our industry is finally getting excited about the built world again, we shouldn’t waste this opportunity - it’s too important that we modernize our industrial base, and robotics must play a role in it.

Subscribe now

My Irish friend Chris was too busy firefighting comms disasters to edit this one, so please bear with me!

Low productivity growth means low economic growth. […] Low economic growth also means the rise of smashmouth zero-sum politics, as gains for one group of people necessarily require taking things away from other people. Zero-sum politics lead to corrosion of the national spirit of opportunity and growth.

Europe’s fundamental values are prosperity, equity, freedom, peace and democracy in a sustainable environment. […] If Europe can no longer provide them to its people – or has to trade off one against the other – it will have lost its reason for being. The only way to meet this challenge is to grow […]

These two quotes seem written by the same author, but they are not. Very different people wrote them: the first one is from Marc Andreessen and Ben Horowits’ “The Little Tech Agenda”, while the second comes from Mario Draghi’s report “The Future of European Competitiveness”.

When people of this caliber not only acknowledge the existence of a problem (zero-sum politics) but also agree on the cause of that problem (lack of growth), then we need to look deeper into what’s happening.

In the last years, I’ve grown more and more concerned about the polarization of our societies. I’ve asked myself what causes a country or a continent to polarize the way the Western World has done. However, I realize now that polarization is impossible to solve without addressing the real root cause, which is zero-sum politics.

A perfect society does not exist, because societies are not static. There is always something to fix, something to change, because the needs of the people living in said society change with time. Fixing and changing things usually requires investments, and here is where things break: without growth, giving to someone means taking from someone else. This is the only thing politics can do when the pie stops growing.

As the Latin said “mors tua, vita mea”, your death is my life: this is what causes polarization.

Polarization is bad because it destroys interpersonal trust, and interpersonal trust in a country is shown to be correlated to GDP per capita.

Lack of interpersonal trust is bad, but lack of trust in those who govern might be even worse.

I come from a country, Italy, that has notoriously short-living governments.

Political satire aside, short governments equal instability, and instability means that you can’t rely on political decisions too much.

If I want to invest my time and money in a country, I want to be sure that the rules of the game don’t change 6 months after I have committed to it. Not trusting those who govern the country is notoriously bad for business, but when societies are polarized this is even worse.

Ideally, the outcome of a democratic parliamentary process should be laws to which both the majority and the minority have contributed. When the governments that follow one another are expression of diametrically opposed views, however, the newly elected government will have to do what it promised to its electors - destroy everything the previous one did.

Isn’t it what it is happening right now in the US? I guess it’s not by chance that trade uncertainty has soared since elections in 2024.

Finally, polarization is bad because instills a “my side vs your side” mentality - which makes us overlook bad things when done by our side

And the villains have a lot of practice using normal people’s flaws to divide them. When society is politically divided, power-hungry people exploit those divisions to rise to the top — we tell ourselves “He may be a bad guy, but he’s a bad guy on my side.” At least Hitler will fight the communists; at least Stalin will keep the capitalists at bay.

Noah Smith

This mentality is very dangerous, because “villains” have no loyalty, no ethics, and no friends, they only have their bank account and hunger for power.

Politically divided societies are the primordial broth of dictatorships and wars - and we are getting a taste of it.

Coming back to the original question, how do we prevent all of this? The best way to solve this mess is growth. When the pie is growing, everyone can have it’s piece and more.

Growth is hard, and requires long-term planning.

How do we make sure European companies don’t overpay for energy?
How do we bring back manufacturing on our soil, when China has become so efficient in producing things?
How do we deal with tariffs without turning our best (and most powerful) ally into an enemy?
How do we make the workforce more efficient to cover for the aging population? How do we build stuff when we don’t control the supply chain of the materials needed to build stuff?
How do we build enough chips for AI to fulfill its promises?

I think these questions make current politicians uncomfortable because they require long-term planning, and polarized audiences are not receptive to long-term plans.

Without being too naive, and keeping in mind that investors are in this game to make money and are not philanthropists, I still think the private sector could do better in allocating capital.

Do we really need another AI website maker, or chatbot for customer support, or automated SDRs? Could we take some time to understand which technologies will be absolutely essential in the future, and allocate capital towards solving these big problems?

I think we can, and we should.

Subscribe now

My long-haired companion Chris edited this piece by rolling back the clock almost three decades to the explosive and psychedelic big beat of “Dig Your Own Hole”.

Each time I read a piece by Ben Thompson, I wonder why he’s not more famous than other well-known bloggers (in my bubble).

I’ve spent much time recently reading about China, clicking on every single link in Kyle Chan’s pieces and falling into a very deep rabbit hole. Most of Kyle’s work is about how China managed to become the global manufacturing powerhouse that it is today, and how Western countries enabled that.

What I could not understand was what it implied for the US and Europe today, and how the implications would influence industrial policies over the next few years.

Then, I stumbled across this piece that Ben Thompson wrote shortly after the 2024 US election. Reading “A Chance to Build” made things click.

The fact is that we didn’t get where we are today by chance - the current situation is one that we designed on purpose. For a long time, because of globalization, the West didn’t need more industrial output.

A problem of atom-based products is that they have non-zero marginal costs: producing things has a cost, like the materials used and labor. For every additional thing you sell, the cost to produce it doesn’t materially improve. Software is different, and that’s why it scales so well - you apply some brainpower, but the cost to produce it doesn’t change with how many licenses you sell. As the world started to open up, companies in the West realized that people in the East were willing to work for less, and offshored any process that required a lot of work and little intellectual property - Fairchild Semiconductors opened their first semiconductor assembly facility in Hong Kong, where wages were $0.25 /hour, in 1964.

Previously, hardware had been the most valuable product while software was seen as a commodity. Once the intellectual property rents of software had been separated from the capital costs of hardware, however, it was inevitable that software companies would trade at higher valuations, attract more investment, and come to dominate the U.S. technology industry, while hardware companies withered. The advent of the internet dramatically accelerated these trends, enabling further separation of software from hardware via the cloud.

https://americanaffairsjournal.org/2021/08/the-value-of-nothing-capital-versus-growth/

When we look at the value chain of many technological products today, we see the same pattern: iPhones are designed in Cupertino but assembled in China, Nvidia’s GPU are designed in Santa Clara but taped out in Taiwan, more than half Teslas sold in the world are designed in Austin but assembled in China, same goes for many other products.

This approach worked really well both for companies and consumers. Businesses have leveraged all the value created by software to grow their stock value, meanwhile customers live in a service and intellectual property economy with higher wages and better working conditions.

Our economies just didn’t need more industrial output, and the result is clear in the graphs below: in the last 20 years West’s share of manufacturing value added has constantly declined in favour of China, as well as the energy generated.

To many, this looked like a bargain.

We could finally stop working in factories, and start working in offices. “Building” moved from meaning turning raw material into objects to meaning turning source code into executable programs.

The problem is that, as Kyle Chan puts it, this is a Faustian bargain, and now we are starting to see the effects of it.

It’s way easier to start with labour intensive and IP-light products before moving upstream in the value chain, than it is to do the opposite. This is exactly what China did.

China never lacked hands or brains, it lacked know-how, and built a strong strategy to acquire it.

How this happened is outside of the scope of this post, and very far from being my expertise - let’s just say it has a lot to do with a very articulated industrial policy, heavy investments in commodities, short-sighted strategies from European/American companies, far-sighted strategies from Chinese companies, and technology spillover.

Now here’s the issue: as we’ve said before, it’s easier to move up the value chain once you control the bottom of it than it is to move back down the value chain once all you have is know-how.

China now dominates the production of most commodities needed to build basic hardware on which Western software should run. Moreover, there are plenty of skillful hands that know how to turn these materials into hardware, something which Europe has been gradually losing.

Jobs went on to urge that a way be found to train more American engineers. Apple had 700,000 factory workers employed in China, he said, and that was because it needed 30,000 engineers on-site to support those workers. “You can’t find that many in America to hire,” he said.

“Steve Jobs” by Walter Isaacson

Now China has a unique opportunity that it has created with its own hands and is working hard not to screw up: moving up the value chain while controlling the foundations of it.

The West has slowly moved up the value chain, offshoring large volume, low value added parts of the process to focus on highly specialized, low volume and high value added ones. Now that China is threatening the apical position with its products, it’s very painful to turn back and reconquer the foundational blocks.

When Mr. Melo bought Caldwell in 2002, it was capable of the high-volume production Apple needed. But demand for that had dried up as manufacturing moved to China. He said he had replaced the old stamping presses that could mass-produce screws with machines designed for more precise, specialized jobs. Mr. Melo thought it was ironic that Apple, a leader in offshore manufacturing, had come calling with a big order. “It’s hard to invest for that in the U.S. because that stuff is purchased very cheaply overseas,” he said.

https://www.nytimes.com/2019/01/28/technology/iphones-apple-china-made.html

On the other hand, once those owning the know-how at the top of the value chain need your products desperately, you can force them to trade them for some of their know-how. This has enabled China to move up in the value chain and slowly replace Western products with their products.

This is the story of Apple up-skilling their suppliers, that in turn start supplying local brands that end up competing with Apple’s products. This is the story of Tesla, that has taught Chinese brands how to build highly automated, vertically integrated factories, only to see the very same Chinese brands replace Tesla in markets like Europe. This is the story of Alstom and Siemens, that first exported the designs of their trains to China, and then were forced to merge to withstand the competition of the very same Chinese companies they helped emerge.

So what happens when borders close instead of opening, tariffs are imposed instead of being removed? For a long time software was the name of the game, because we took hardware for granted. We packed our homes, phones and cars with software because we never doubted we would have enough hardware or energy to run it. Now that the Pax Americana seems to be over, we are moving towards a world in which the most important technologies are made of atoms, not bits - and we need to relearn how to turn atoms into objects if we want to stay relevant.

This trend applies to the West as a whole, but more so to Europe, as we became particularly vulnerable on the topic.

First, Europe is extremely globalized.

This failure to lower internal barriers has also contributed to Europe’s unusually high trade openness. Since 1999, trade as a share of GDP has risen from 31 per cent to 55 per cent in the eurozone, whereas in China it rose from 34 per cent to 37 per cent and in the US from 23 per cent to just 25 per cent. This openness was an asset in a globalising world. But now it has become a vulnerability.

Mario Draghi

This means that we have double to lose because not only does trade make up a good part of our companies’ revenues, but also we rely on trade for a good part of our needs.

The solution to this is as easy in theory as it is hard in practice: remove the internal barriers and make trading among European countries much cheaper than it is to trade with the rest of the world.

Second, Europe is not energy independent.

The US has extensive shale oil and gas reserves, which had been inaccessible until recently, but the shale oil revolution radically changed the global energy landscape from 2010. US production of both natural gas and crude oil has more than doubled in just over a decade, providing abundant and cheap energy, and turning the US into a net energy exporter. In 2010, the US had a trade deficit in energy worth $360 billion, which by 2022 had turned into a surplus of $65 billion.

https://www.cer.eu/insights/why-europe-should-not-worry-about-us-out-performance

It is pointless dedicating any words on acknowledging the problem of being unable to produce enough of our own energy. As Dealroom puts it in his latest report Accelerating Europe: “Shale boom gave the US economy a $0.5 trillion boost, plus a strategic advantage.”

It is of utmost importance that we roll our sleeves and we diversify our energy mix, using all the resources we have (yes, nuclear, I’m looking at you) and investing in companies innovating in the energy sector - otherwise we will never be competitive again.

Third, European population is aging and declining.

There is little tech and venture capital can do to revert this trend. What can be done is, however, empower the few workers left with tools to become more productive.

Europe has lost control of many commodities that are needed to build things: we don’t produce silicon wafers, we produce little steel and aluminum, and same for many other commodities. The only way to get back into the game is to innovate on the cost side, for example finding more energy-efficient processes as energy is the main cost driver in many cases.

We also lost the ability to produce things efficiently at high volumes. We often see ourselves as artisans, but artisans don’t scale: we need automation and we need optimization, as soon as possible. Long live the artisans crafting shoes, but let’s use robots to build cars.

As Ben Thompson concluded in his piece: “It’s Time to Build.”

Geopolitical turmoil is dangerous but creates opportunities - if there is a moment to disrupt otherwise slow moving industries, that moment is now.

The real opportunity is taking advantage of robotics and AI to make physical goods into zero marginal cost items in their own right (outside of commodities; this is what has happened to chips: assembly and testing are fully automated, which makes a U.S. buildout viable).

https://stratechery.com/2024/a-chance-to-build/

I’m excited to see momentum picking up in Europe, companies building in the right spaces and funds investing in the right companies.

When we stop idolising the US as the prom queen, when we motivate young engineers to be excited about the built world, when we treat our neighboring European countries as partners and not competitors, only then will we be able to seriously tackle the fundamentals problems that threaten the very future of our democracies.

C’mon Europe, it’s time to build!