Artificial intelligence has become the market’s favorite story, but one of Wall Street’s most famous skeptics is warning that the script will not have a happy ending. Michael Burry, the “Big Short” investor who called the housing crash, is arguing that the current AI boom is structurally flawed and destined for a painful reset, and he is using Warren Buffett’s own playbook to make his case. By contrasting Burry’s alarm with Buffett’s quieter, more selective bets, I see a roadmap for how this bubble could burst and what might be left standing afterward.
Burry’s AI bubble thesis: trillions in spending, thin real-world payoff
Michael Burry’s core argument is that the AI trade has sprinted far ahead of the underlying economics, with investors extrapolating early excitement into permanent, sky-high profitability. He has warned that AI-related earnings are being flattered by unusual conditions, including customers that are effectively subsidized by their suppliers and hyperscalers that are spending aggressively without a clear line of sight to durable returns. In one detailed breakdown shared with value investors, Burry was cited as arguing that AI-linked profits are inflated and that the true economics, once subsidies and one-off deals fade, will look far less impressive than the market currently assumes, a view that has been dissected in value investing circles.
His concern is not just about frothy valuations, it is about the scale and direction of capital being poured into the theme. Burry has pointed to what he describes as “trillions of dollars of spending with no clear path to utilization by the real economy,” arguing that this wave of investment risks becoming a misallocation if it fails to translate into broad productivity gains or defensible profits. In his telling, AI infrastructure and chips may be selling briskly today, but if customers cannot convert that capacity into sustainable cash flow, the current boom will look like a classic capital cycle blowoff, a warning he has tied directly to the surge in demand for the company’s pricey, powerful chips that dominate the narrative in AI hardware.
The escalator in Baltimore: how Buffett’s story shapes Burry’s view
To explain why he thinks AI will disappoint investors even if the technology works, Burry has reached for an old Warren Buffett anecdote about a department store in Baltimore. In that story, Buffett described installing an escalator that delighted shoppers but did little for the store’s bottom line, because the benefit accrued almost entirely to customers rather than shareholders. Burry has seized on this as a metaphor for AI, arguing that if the gains from smarter software and automation mostly flow to users in the form of lower prices or better experiences, then the companies footing the bill for AI infrastructure may struggle to earn attractive returns, a point he has linked explicitly to the way value accrued in the escalator example.
In Burry’s reading, the biggest surprise would not be that AI fails technically, but that it succeeds and still leaves investors underwhelmed because competition drives away excess profits. Unless a company can charge monopoly-like rents, he argues, the pattern from the Baltimore escalator repeats: customers win, margins compress, and the capital providers who funded the build-out are left with mediocre returns. That logic underpins his skepticism toward the idea that every business buying AI tools will magically unlock a durable competitive advantage, a narrative he has challenged in conversations about how AI implementation will actually play out in the real economy, including his recent comments about how adoption could mirror that early escalator rather than a gold mine for shareholders, as recounted in his discussion of Buffett’s story.
From theory to trade: Burry’s bets against the AI darlings
Burry is not leaving his AI skepticism at the level of theory, he is expressing it in his portfolio. After spending a previous quarter in a rare risk-on stance, building long equity exposure and call options across healthcare, consumer, and technology names, he has shifted back toward a more defensive, contrarian posture. Analysts tracking his positions have described this as a “different kind of contrarian bet,” with Burry leaning into the idea that markets are overpaying for fashionable growth stories and underpricing the risk that AI enthusiasm fades, a stance that has been highlighted in reviews of his stock portfolio.
The most visible expression of that view is his move to bet directly against Nvidia, the chipmaker that many see as the purest play on the AI build-out. Burry has described Nvidia as dangerously reliant on hyperscaler spending and has said, “I do not see how that math works,” signaling that he believes current expectations for growth and profitability are unrealistic given the cyclical nature of capital spending by big cloud providers. By shorting what he calls the “purest play” on the AI chip boom, he is effectively wagering that the market has misread the durability of today’s demand spike and that a prolonged slump is coming once spending normalizes, a thesis he has laid out in detail in his critique of Nvidia’s prospects.
What Buffett is actually doing with AI in Berkshire Hathaway’s portfolio
Warren Buffett’s own portfolio tells a more nuanced story about AI than the caricature of a technophobe sitting out the boom. Through Berkshire Hathaway, Buffett has significant exposure to companies that are using AI as a tool rather than selling it as a product, with a large share of his capital tied up in businesses that benefit from digital infrastructure and data-driven services. One breakdown of Berkshire Hathaway’s holdings shows that 33.5% of Warren Buffett’s 304 billion dollar portfolio is invested in four artificial intelligence related stocks, underscoring that he is not ignoring the theme but is choosing his spots carefully within a concentrated set of AI stocks.
Heading into 2026, Buffett has roughly 23% of Berkshire Hathaway’s portfolio invested in three artificial intelligence related stocks, a figure that reflects both his comfort with certain tech-adjacent franchises and his reluctance to chase the most speculative names. Rather than piling into the hottest chip designers or unproven software platforms, he has favored companies with entrenched customer bases and strong cash flows that are integrating AI into existing services, a pattern that can be seen in the way Berkshire Hathaway’s capital is allocated across its portfolio invested in Artificial Intelligence Stocks Heading Into the new year.
How Buffett’s discipline reinforces Burry’s warning
When I put Burry’s alarm next to Buffett’s positioning, the contrast is striking but not contradictory. Burry is loudly warning that the AI boom will end badly, arguing that the market is treating every AI narrative as if it were a guaranteed moat, while Buffett is quietly proving that even a legendary long-term optimist will not pay any price for a fashionable story. In a recent conversation with Wall Street professionals, Burry raised the alarm on an AI bubble and suggested that a prolonged slump is coming, even as he acknowledged that some companies will find real, lasting advantages from the technology, a nuance that came through in his comments about how only a subset of firms will achieve a true competitive advantage.
Buffett’s own remarks and portfolio choices effectively validate Burry’s escalator analogy, even if the two men are not coordinating their messages. By focusing Berkshire Hathaway’s AI exposure on a handful of durable franchises and avoiding the most hyped chip and software names, Buffett is behaving as if he expects AI to be transformative for customers but only selectively lucrative for shareholders. Burry, for his part, has used that same Baltimore escalator story to argue that the biggest surprise would be a world where AI works technically yet fails to deliver the profits investors are baking in, a perspective he has reiterated in interviews where he, as the investor of Big Short fame, took aim at the AI boom in conversations with figures such as Clark and podcaster Dwarkesh Patel, as recounted in coverage of his AI warnings.
More from Morning Overview