Category: Uncategorized

By Juan Garces De Marcilla, Founding Member of Montesa Capital

There are two AI debates happening right now, and almost everyone is having only one of them. The first — whether AI stocks are overvalued in the next 18 months — is the one filling earnings calls and financial media. The second is the one almost nobody outside this firm is asking out loud: whether AI represents a phase shift within capitalism, the way the Information Age was a new stage within capitalism, or a genuine mode-of-production shift — the kind of structural break that took the world from feudalism to capitalism in the first place. That’s the question this piece takes on. We can ask it because Montesa Capital doesn’t treat economics as a single school of thought to defend. We take Keynes seriously on demand, Hayek seriously on prices and information, and Marx seriously on the mechanics of how value gets produced and captured — and we think the answer to where AI is taking the economy requires all three.

The Mechanism, in Two Languages

Standard economics models economic output as a function of two inputs: capital and labor (formally, Y = A·K^α·L^(1−α), the Cobb-Douglas production function — don’t worry about the notation, the point is simple: output depends on machines and people, not machines alone). If AI drives labor’s contribution toward zero, you are not just changing who has a job. You are changing one of the two variables the entire system runs on.
Marxian economics has a sharper way of naming what happens next: a realization crisis. Capital can produce a commodity, but it can only “realize” — actually collect — the value of that commodity if someone, somewhere, has the income to buy it. In plain terms: it doesn’t matter how cheaply or efficiently you can build something if nobody can afford to buy it. Roughly 68–70% of U.S. GDP is consumer spending. If AI hollows out wage income broadly enough, you don’t just strand workers — you eventually strand the capital that depends on those workers’ paychecks to generate a return. This isn’t a fringe theory; Marx anticipated a version of this scenario directly in an 1858 manuscript known as the “Fragment on Machines,” describing an economy approaching full automation as one where the entire logic of labor-driven value starts to break down.
To be clear about scope: nothing in the verified labor data points to this happening fast or completely. The most aggressive credible estimates — Goldman Sachs, MIT’s Daron Acemoglu — put workforce displacement in the high single digits over a decade, not a collapse in 18 months. This is a long-horizon structural question, not a near-term catalyst. But “long-horizon” doesn’t mean “not real” — it means the market isn’t pricing it, because markets are bad at discounting diffuse, decade-scale risks. That mispricing is exactly why it’s worth thinking through now.

Candidate Paradigm #1: Technofeudalism

The most prominent answer to “what comes after labor-driven capitalism” comes from economist and former Greek finance minister Yanis Varoufakis, in his 2023 book Technofeudalism: What Killed Capitalism. His argument: the defining features of capitalism — competitive markets and profit — have already been displaced by something closer to feudal rent. Platforms (Amazon, Google, the major cloud infrastructure providers) don’t compete in markets; they own the “land” — the digital infrastructure — and charge rent for access to it. Varoufakis estimates Amazon extracts roughly half of every dollar a third-party seller earns on the platform, which he likens to a feudal lord’s ground rent rather than a competitor’s profit margin. In his framing, even traditional capitalists become “vassals” — dependent on cloud infrastructure they don’t own or control, the same way medieval lords depended on land they held at the pleasure of a king.
In plain terms: instead of “the market decides who wins,” it becomes “whoever owns the pipes decides who gets to play, and takes a cut regardless of whether you win.” That’s a real and observable shift in how value is captured, even if you don’t accept Varoufakis’s stronger claim that capitalism is already fully dead. AI compute access — who controls the chips, the data centers, the foundation models — is arguably the newest and most extreme version of this dynamic: a tiny number of “cloudalists” control the infrastructure every other business increasingly has to rent in order to function

Candidate Paradigm #2: The Resource-Based Economy

The other serious answer — less mainstream in finance circles, but worth a fair hearing — comes from Jacque Fresco’s Venus Project and its associated Zeitgeist Movement: a resource-based economy (RBE), in which goods and services are allocated without money, credit, or debt, with resources treated as the common heritage of all inhabitants and distribution coordinated by technology rather than price.
The standard economic objection to this is Friedrich Hayek’s “knowledge problem” (from his 1945 paper “The Use of Knowledge in Society”): prices aren’t just cost tags, they’re a real-time information system that aggregates millions of people’s local knowledge and preferences. No central planner — human or algorithmic — has historically been able to replicate that without some price-like signal, and the most-cited failure case is Soviet “War Communism” (1918–1921), Lenin’s own attempt to abolish money and market exchange, which collapsed into shortages and spontaneous black markets within three years.
In fairness to the RBE position, that historical case is more confounded than it’s usually given credit for. War Communism wasn’t a clean experiment — it was implemented in the middle of a civil war, under an Allied economic blockade, using forced grain requisitioning (prodrazvyorstka) that destroyed farmers’ incentive to produce regardless of the money question, on top of infrastructure already shattered by World War I, using paper ledgers and telegraphs as the only available coordination technology. Attributing its failure purely to “moneylessness doesn’t work” overstates what the historical record can actually prove — it’s at least as plausible that war, coercion, and the total absence of any real-time information infrastructure did the damage, independent of the price question.
That matters because the RBE argument today isn’t “trust a Soviet bureaucrat to guess what people want.” It’s “let an AI system with real-time visibility into actual consumption, logistics, and inventory data allocate resources without needing a price signal to do it.” That’s a meaningfully different proposition than what failed in 1918–21, and modern logistics already does a limited version of this inside capitalism — Amazon’s own inventory and demand-forecasting systems route goods without waiting for price signals to clear at every node. Whether that scales to replacing markets entirely, rather than optimizing within them, remains genuinely unresolved — it is not disproven by the Soviet case, but it isn’t proven by Amazon’s warehouse algorithms either. This is a live, unsettled question, not a closed one.

What This Means, Practically

Both paradigms describe a world where traditional wage-labor stops being the organizing principle of the economy. They differ entirely on what replaces it: technofeudalism keeps private ownership and exchange but concentrates power in whoever owns the infrastructure (rent over profit); the resource-based economy removes price and ownership altogether in favor of algorithmic allocation. A third, less radical path — favored by most mainstream economists who take the labor-displacement problem seriously, from Daniel Susskind to various UBI proposals — doesn’t change the mode of production at all; it just redistributes capital’s gains back to displaced labor through taxation and transfers, leaving markets and money fully intact. That third option is the status quo’s self-correcting answer, and it’s the one with by far the most real-world precedent (it’s just a bigger version of unemployment insurance and progressive taxation).
For capital allocation today, the practical takeaway isn’t to bet on any one of these outcomes — it’s to recognize that the infrastructure layer (chips, cloud compute, data centers, energy supply) is the asset class that wins under either of the two more radical paradigms, technofeudal or resource-based, because both require someone to own and operate the physical and digital “pipes” regardless of how the output gets allocated. Betting on the rent-collecting layer of the AI economy is a hedge that works whether the next fifty years look more like Varoufakis’s cloudalism or like a more redistributive patch on capitalism as we know it. The one scenario it doesn’t hedge well against is the full resource-based economy outcome — which, for now, remains the most speculative and least empirically grounded of the three.

Montesa Capital research notes are for informational purposes and do not constitute investment advice.

By Juan Garces De Marcilla, Founding Member, Montesa Capital

The AI bubble debate gets confused because two different questions are being asked at once: is AI overvalued eventually, and is it about to crash soon. The first question is genuinely unsettled — reasonable, credentialed people disagree, and the structural risk we wrote about previously (AI’s success slowly eroding the consumer income it depends on) is real but plays out over a decade, not over the next earnings cycle. It tells you nothing about whether 2026 or 2027 brings a crash.
The second question — is a burst likely in the next 18 months — is the one that actually matters for capital allocated today. Here is the strongest verified case that it isn’t.

1.This isn’t 1999.The revenue is real.

The dot-com comparison breaks down on contact with the numbers. Microsoft’s AI business has surpassed a $37 billion annual revenue run rate, up 123% year over year, with commercial backlog standing near $627 billion. Nvidia’s Q4 data center revenue hit $62.3 billion, up 75% year over year. These are not projections or pre-revenue startups burning venture capital on Super Bowl ads — they are large, profitable, audited businesses with accelerating cash earnings. Dan Ives, the Wedbush analyst who has covered tech since the dot-com era, makes the contrast directly: in 1999 the average tech stock traded at 30x revenue on little more than slide decks; today’s AI leaders are generating hundreds of billions in cash against real infrastructure and real customers.

2. Adoption is still early — the demand runway hasn’t been tested yet.

Bubbles burst when the market for a product turns out to be smaller than investors assumed. That’s hard to argue convincingly here: by Wedbush’s count, only about 3% of US companies and less than 1% globally have meaningfully deployed AI. Whatever you think of current capex levels, the claim that AI has already saturated its addressable market and is now running out of buyers doesn’t survive contact with the adoption data. There is a lot of room left to grow into the spending before demand becomes the limiting factor.

3. Balance sheets went into this cycle unusually strong..

This is the most underappreciated point in the debate. Unlike 2000 — and unlike a classic credit bubble — most hyperscalers entered the AI buildout with historically low leverage or outright net cash positions. By late 2025, the aggregate liabilities-to-assets ratio across major hyperscalers had fallen to around 48%, near 2015 lows, compared with roughly 80% for the S&P 500 broadly. That matters because it means even as capex begins to outpace operating cash flow — which independent analysis projects could happen in aggregate by roughly Q3 2026 — these companies are financing the gap from a position of strength, with investment-grade balance sheets, not from a position of fragility.

4. Independent, mechanical frameworks back this up.

JPMorgan ran the AI rally through a formal five-factor bubble-diagnostic test in December 2025 and concluded that current investment is linked to actual enterprise revenue rather than speculation alone — a structured, falsifiable test, not a sentiment-based opinion. Separately, Federal Reserve Chair Jerome Powell has drawn an explicit distinction from the dot-com era on the same grounds: that AI companies are generating real revenue and that data center spending is contributing to measurable GDP growth, not just to stock multiples.

The one thing actually worth watching

Intellectual honesty requires naming the real near-term risk rather than ignoring it: the financing structure of this buildout is shifting from cash flow to debt faster than the bulls like to admit, and that shift is not evenly distributed. Oracle is the clearest stress case — its five-year credit default swap spread roughly quintupled between early 2025 and March 2026, reflecting real concern about its debt-funded, highly customer-concentrated buildout tied largely to OpenAI. That is a legitimate watch item. But it is a concentration risk in specific balance sheets, not evidence of a sector-wide bubble about to pop. The distinction matters: a single overleveraged counterparty stumbling is a credit event, not a verdict on the asset class.

The bottom line

The strongest version of the bear case isn’t “AI demand doesn’t exist” or “the revenue is fake” — both claims are contradicted by the earnings data. The strongest version is a slower, structural one: whether the income AI displaces over the next decade quietly narrows the customer base that all of this revenue ultimately depends on. That’s a real question, and we’ve made the case for it elsewhere. It is not, however, an 18-month catalyst. For the time horizon that determines whether capital deployed today survives the next two years, the data argues for real revenue growing into real spending — with leverage concentrated in a small number of names worth watching individually, not feared collectively.

Montesa Capital—research notes are for informational purposes and do not constitute investment advice.

By Juan Garces De Marcilla

There is a particular kind of investor who spends enormous energy appearing sophisticated. They speak in factors and alphas. They diversify across asset classes with the confidence of someone who has confused complexity with wisdom. They would never buy Tesla — too retail, too Elon, too obvious. They would never buy Bitcoin — too volatile, too speculative, too unsophisticated for someone of their analytical caliber.

They are, in many cases, dramatically underperforming the person who simply bought both and went to the beach.

I want to challenge a misconception that runs deep in professional investing circles — the idea that sophistication is correlated with returns. It isn’t. In fact, sophistication is often the enemy of returns. It gives intelligent people permission to overcomplicate decisions that are, at their core, straightforward. The most dangerous investor in the room is rarely the one who knows too little. It is the one who knows just enough to find a counterargument for every obvious opportunity — and who has mistaken the ability to complicate a decision for the ability to make a good one.

The Data Doesn’t Lie

In 2024, the S&P 500 returned approximately 23%. Bitcoin returned roughly 120%.

The world’s most sophisticated investors — hundreds of PhDs, billions in infrastructure, decades of institutional experience — collectively failed to beat a simple index fund that a 22-year-old could have bought in five minutes. The top traditional hedge fund by percentage return was Discovery Capital at 52.7%. Most of the industry’s biggest names returned between 15% and 22%. The average hedge fund returned 8.3%.

The funds that did outperform didn’t do it through complexity. D.E. Shaw’s best performing fund returned 36% through concentrated macro conviction. Citadel’s highest returning strategy was its most directional. The exceptions prove the rule — the path to outperformance in 2024 ran through clarity and conviction, not sophistication and diversification.

This is not an anomaly. It is a pattern. And it has a name — the complexity premium in reverse. The more moving parts, the more ways to be wrong.

The Microsoft Lesson

In 1986, Microsoft went public at $21 per share. It was obviously the operating system of the personal computing revolution. It was not a secret. It was not a contrarian bet. It was, to anyone paying attention, the most straightforward investment available. Ten thousand dollars invested at that IPO would be worth approximately $40 million today.

You did not need to be a genius. You needed to be paying attention, have conviction, and resist the temptation to do something more interesting with your money.

This is the pattern that repeats across every major wealth creation event in market history. The opportunity is usually visible. It is usually obvious in retrospect and merely uncomfortable in the moment. The barrier is not analytical horsepower. It is the psychological courage to act on simple conclusions in the face of social pressure to appear more sophisticated. Every generation gets one or two of these moments. The tragedy is how many people are too busy being clever to notice them.

The Optimus Problem

I I believe we are in one of those moments right now.

Tesla’s Optimus robot is, in my view, one of the most asymmetric investment opportunities in the history of public markets. Consider what it actually is — a general purpose humanoid robot, manufactured at scale, by arguably the most advanced manufacturing operation on earth, designed to perform physical labor across every industry simultaneously. The global market for physical labor runs into the tens of trillions of dollars annually.

I may be wrong about Optimus. The history of robotics is littered with products that looked transformative in controlled conditions and failed in the complexity of the real world. The competitive landscape is real — capable and well-funded companies are racing toward the same prize. There is no guarantee Tesla wins.

But here is the asymmetry that matters: the downside of being wrong is owning a position in the world’s most valuable car company with some of the most advanced energy and AI infrastructure on earth. The upside of being right is owning a meaningful stake in what could become the largest product market in human history. That gap between downside and upside is what investing is actually about. You do not need certainty. You need the gap to be wide enough that being right occasionally is sufficient.

The sophisticated investor looks at Tesla and sees a volatile, over-hyped consumer brand with an eccentric founder. I look at it and see a manufacturing and robotics platform that the market is still, somehow, undervaluing on the scenario that matters most.

The Bitcoin Corollary

The same principle of asymmetry applies in crypto, but with an important inversion. Here, sophistication tends to make things worse rather than better.

The market is littered with alternative cryptocurrencies — thousands of them — each with a white paper, a use case, a community of believers, and a price chart that on closer inspection correlates almost perfectly with Bitcoin’s in every meaningful downturn. This is not a coincidence. It reflects a structural reality the market has repeatedly confirmed: when Bitcoin falls, almost everything falls with it regardless of merit.

This creates a specific problem for the altcoin investor worth stating plainly. Every alternative cryptocurrency carries two distinct risks. The first is Bitcoin risk — the possibility that the entire asset class reprices downward, taking everything with it. The second is idiosyncratic risk — the possibility that the specific token you chose turns out to be technically unsound, outcompeted, or simply a fraud. Some altcoins have outperformed Bitcoin over specific periods — Ethereum and Solana are legitimate examples. But they are the exceptions in a landscape where the vast majority of tokens have gone to zero or close to it.

The question is not whether altcoins can outperform. Some do. The question is whether the expected value of picking correctly — in a market designed to obscure quality, riddled with information asymmetry, and dominated by insiders — justifies the additional layer of risk for a non-specialist investor.

For most people it does not. Bitcoin has one job. It has been doing that job for fifteen years with a track record no alternative can match. Adding complexity to that thesis does not make it more sophisticated. It makes it more fragile. The burden of proof sits with the altcoin, not with Bitcoin. And in my experience, that burden is rarely met with anything more than a compelling story.

What This Actually Requires

I am not arguing for laziness. Simplicity is not the same as ignorance. There is a profound difference between the investor who buys Tesla and Bitcoin because they have done the work and arrived at a clean conclusion, and the investor who buys them because someone on the internet told them to. The former is one of the hardest things to do in investing — sitting with a simple thesis under enormous social pressure to complicate it, holding through volatility that would shake most professionals loose, and maintaining conviction when the consensus shifts against you.

None of that is easy. But none of it requires a PhD, a Bloomberg terminal, or a forty-page discounted cash flow model. It requires clarity about what you actually believe, honesty about what you don’t know, and the discipline to act accordingly.

The Larger Point

The next decade will produce extraordinary wealth creation. Artificial intelligence, humanoid robotics, decentralized systems, biological science — these forces are not arriving quietly. They are restructuring the global economy in real time, compressing decades of change into years, and creating mispricings that will look, in retrospect, as obvious as Microsoft in 1986.

Most professional capital will miss the majority of it. Not because the managers aren’t intelligent — many of them are extraordinarily so. But because their intelligence is deployed in the service of complexity, and complexity is the enemy of the clear-eyed bet. The 2024 data makes this not a theory but a fact. The best performing funds were the most convicted. The most diversified were the most average.

The opportunity is visible. It is sitting in plain sight. The only question is whether you have the clarity to see it and the conviction to act on it before consensus forms.

You don’t have to be smart. You just have to be willing to act on what you can already see.