Warren Buffett cash pile and Michael Burry AI short: what two contrarian investors are signaling

Berkshire Hathaway’s record $397 billion cash pile and Michael Burry’s AI short are not the same trade, but they point in the same direction. One says capital looks hard to deploy at attractive prices. The other says the AI winners may already be priced for a future that assumes too much, too soon.

That is the real tension behind the Warren Buffett cash pile and Michael Burry AI short story. Berkshire’s cash, which Bloomberg reported earlier this month, climbed to $397 billion in the first quarter after the conglomerate sold a net $8.1 billion of equities. At the same time, JPMorgan described in January how Burry had roughly $1 billion in short positions against AI stocks, a setup it compared with the run-up to the dot-com crash.

The overlap is worth attention, but it is not proof of a market top. Berkshire’s cash hoard is a fact. Reading it as a direct verdict on AI valuations is the leap.

What Berkshire’s record cash hoard actually says

The cleanest read on Berkshire’s balance-sheet mountain is also the least dramatic. The company sold more equities than it bought in the quarter, and its cash position jumped to $397 billion, the highest level ever reported by the firm, Bloomberg reported earlier this month. That came during Greg Abel’s first full quarter as chief executive, and operating earnings rose as underwriting improved in Berkshire’s insurance businesses.

So this was not distress selling. It was a healthy business choosing not to force the issue.

That matters because Berkshire’s hoard is easy to over-interpret. Some investors want it to be a grand statement about AI, or about markets generally. The evidence is thinner than that. What the quarter shows, more plainly, is valuation discipline. If nothing clears the bar, cash piles up. Berkshire has taken that route before, and usually only breaks it when prices shift far enough to create real value.

There is another reason the figure deserves attention. It comes from a company that has rarely had trouble finding a home for money when the odds were right. That makes the current restraint more striking than the number itself.

The firm’s first-quarter results were not weak. Bloomberg reported that operating earnings jumped, helped by better underwriting results in the insurance businesses. In other words, Berkshire is not hoarding cash because the engine is broken. It is hoarding cash because the market has not offered enough.

That does not mean Warren Buffett, or Greg Abel, are making a direct call on AI. It could reflect stretched valuations across the board, limited opportunities in Berkshire’s preferred lanes, or a simple preference for liquidity while yields on cash remain decent. But the broad signal is still plain. A company built on deployment discipline is choosing patience over action.

The market likes to treat cash as a metaphor. Sometimes it is just cash. Even so, $397 billion is a hard number to ignore.

Why Michael Burry is short Nvidia and AI stocks

Burry’s case is more pointed, and more useful if the question is AI specifically. In a Substack post last weekend, he called Nvidia “simply the purest play” on AI enthusiasm, according to Business Insider. He also said the company has become “entirely dependent on hyperscaler spending, and I do not see how that math works.”

That is the core of his argument. Nvidia may sell $400 billion of chips this year, while there are less than $100 billion in application-layer use cases to justify that buildout, Business Insider reported earlier this year. The gap is not a rounding error. It is the thesis.

Burry’s choice of target is revealing too. He said betting against Meta would amount to shorting its social media and advertising dominance. An Alphabet short would be a bet against Search, Android, Waymo and the rest of the Google machine. Shorting Microsoft would mean taking on a global office productivity SaaS giant, with Word and Excel tucked inside a business that throws off real cash, Business Insider reported.

Nvidia is different because the bull case is narrower. Business Insider reported that the stock has surged 12-fold since the start of 2023, pushing its market capitalization to about $4.5 trillion. That valuation depends heavily on continued hyperscaler spending. If that spending slows, the cushion is thinner than it looks.

Burry’s skepticism is not limited to Nvidia. He also owns bearish put options on Oracle and has directly shorted the database company in the last six months, extending the same logic to cloud and database infrastructure tied closely to AI demand, Business Insider reported. He has also said he would short OpenAI if it were public, pointing to the ChatGPT maker’s $500 billion valuation, which exceeds the market values of Johnson & Johnson, Bank of America, and Costco, Business Insider reported in October.

That comparison is useful because it shows where Burry thinks the excess lives. Not in AI as a technology. In the price attached to the most crowded names.

Why Buffett and Abel are still sitting on cash

So why does Berkshire’s cash pile matter beyond the headline number? Because it comes from a company that has spent decades proving it can deploy capital when it sees a bargain. That makes restraint more interesting than aggression.

The simplest explanation is still the best one. Berkshire sold $8.1 billion more in equities than it bought during the quarter, and its cash pile rose to $397 billion, Bloomberg reported earlier this month. The firm was also more profitable on an operating basis, helped by stronger underwriting. This was not a business under pressure. It was a business waiting.

That does not mean Buffett and Abel are making a direct call on Nvidia or the rest of the AI trade. It could reflect stretched valuations across the market, limited opportunities in Berkshire’s preferred industries, or simply a preference for liquidity while cash still earns something. But the signal is still useful. A company built on discipline is choosing patience over action.

One caveat is worth keeping in view. Berkshire’s cash hoard is a hard fact; treating it as a specific verdict on AI valuations is an inference. The picture is more nuanced than a simple “smart money is fleeing.”

The AI boom and the 1999 comparison

The dot-com analogy has been dragged out often enough to lose some of its edge, but Burry’s framing gives it a more specific shape. He is not saying AI is fake. He is saying the capital-spending cycle looks stretched relative to the revenue that can support it.

That is where the comparison starts to make sense. Nvidia’s $4.5 trillion market cap, Business Insider reported, implies a level of confidence in future demand that is hard to square with a narrower near-term use case set. When a company may sell $400 billion of chips but the application layer is still under $100 billion, the market is already paying for a lot of future success. That is before the next chip cycle arrives.

Burry’s point about technological obsolescence sharpens the risk. He wrote that Nvidia seems to introduce a new chip solution every year or less, which means today’s hardware can become yesterday’s story faster than investors usually admit, Business Insider reported. In a boom, that pace looks like progress. In a bust, it looks like depreciation with better branding.

Still, the 1999 comparison has limits. The companies funding AI infrastructure today are not dot-com era zombies burning cash with no business model. Microsoft, Google, Amazon and Meta are profitable, cash-generative, and building for reasons that go beyond next quarter’s narrative. They have room to overbuild without immediately falling apart.

That is why the bubble call is harder to make than it was in 1999. The infrastructure may be early. The capital is real. The monetization is the uncertain piece.

Burry has tried to capture that tension with a blunt metaphor. He said AI could be “the anabolic steroid that allows superhuman feats of strength and ability,” but one that could turn out to be “more dangerous than anything,” Business Insider reported. It is a fair line. Steroids do make things bigger. They also tend to distort the system feeding them.

What the market should take from both signals

The useful way to read Berkshire and Burry is not as prophets but as position-takers. Both are declining to buy into the most enthusiastic version of the current market story. That is not a forecast, but it is a clue.

Berkshire’s cash pile says the bar for capital deployment remains high. Burry’s short against Nvidia says the AI trade may be too concentrated in a handful of names whose valuations assume a long runway of spending growth. Put together, they suggest that caution is not a fringe view anymore. It is sitting in plain sight.

That does not mean the AI boom is over. It does mean the easiest money may already have been made in the most obvious winners. The technology can be real and the pricing can still be excessive. Markets have a long, undignified history of getting both things wrong at once.

So the better question is not whether AI matters. It clearly does. The question is whether investors buying the hottest names are paying for a future that still has room to surprise, or one that has already been stretched to the point where disappointment becomes expensive.