AI & ComputeLOT 261122 · 3 min read

AI stock selloff broadens as Burry cites $176 billion cost gap

AI infrastructure stocks slid after Michael Burry renewed warnings about borrowing costs and $176 billion in depreciation costs he says operators may understate through 2028.

By Nathan Brooks665 words

Key takeaways

  1. CoreWeave had fallen 76% from its recent high by Nov. 19, 2026.
  2. On Nov. 19, 2026, Nvidia fell 3.3%, CoreWeave 6.1%, and Nebius Group 9.4%.
  3. Burry estimates AI companies may understate depreciation costs by $176 billion through 2028.
  4. Microsoft posted 48% Azure growth and $18.6 billion in free cash flow in its September quarter, according to Reuters.
  5. Amazon's trailing 12-month free cash flow swung by $57.2 billion, from a $14.8 billion inflow to a $42.4 billion outflow.
Michael Burry Says ‘I Told You So’ as Wall Street Caught Off Guard by AI Selloff
Michael Burry Says ‘I Told You So’ as Wall Street Caught Off Guard by AI Selloff — Akela999 / pixabay

By Nathan Brooks

CoreWeave had fallen 76% from its recent high by Nov. 19, 2026, as investors questioned whether trillions of dollars in AI infrastructure spending would produce adequate returns.

The Nov. 19 losses were severe even before the longer decline. Nvidia fell 3.3%, CoreWeave 6.1%, and Nebius 9.4%. Selling has spread from semiconductor manufacturers to companies that build or rent AI data centers, and even strong earnings reports have struggled to restore confidence.

What did Burry warn investors about?

On Nov. 19, hedge fund manager Michael Burry posted on X to cite earlier warnings about Nvidia, CoreWeave, Nebius Group, and semiconductor stocks. Burry, famous for predicting the 2008 housing collapse, wrote: “I tried to warn you all,” adding, “I didn’t say short. But I provided the information one would need to sell.”

The post landed during a much broader reversal. Goldman Sachs’ Delta-One trading desk called the move a “pretty violent liquidation,” noting that stocks fell sharply from the U.S. opening despite catalysts that initially appeared supportive. The desk said Wall Street’s surprise concerned the broader selloff, not Burry’s predictions specifically.

For investors who spent the past year watching AI shares climb, the reversal has been painful. For Burry, it looks remarkably familiar.

Declines from recent highs as of Nov. 19, 2026:

  • Nvidia (NASDAQ: NVDA): 39%
  • CoreWeave (NASDAQ: CRWV): 76%
  • Nebius Group (NASDAQ: NBIS): 66%
  • VanEck Semiconductor ETF (NASDAQ: SMH): 33%

Can AI infrastructure economics withstand faster depreciation?

Burry’s central concern is not that artificial intelligence will fail. It is that the economics supporting its infrastructure boom may prove unsustainable.

Companies are spending hundreds of billions on specialized chips with uncertain useful lives. If newer processors make older equipment less competitive, data center operators could face depreciation expenses that exceed expectations.

Burry has argued that AI companies may understate depreciation costs by $176 billion through 2028. That $176 billion is his estimate, not a confirmed industry-wide charge.

Financing turns the accounting risk into a commercial risk. CoreWeave and Nebius rely on expensive infrastructure to serve AI customers, while rising borrowing costs could weaken their models if demand or GPU rental prices weaken. Nvidia faces the same chain from the other side: its customers must earn enough from their investments to keep buying new processors.

Does the selloff prove the bubble?

Falling stock prices do not prove that the underlying businesses are deteriorating. The selloff offers a verdict; the filings do not.

Microsoft reported 48% Azure growth in its September quarter while generating $18.6 billion in free cash flow, according to Reuters. That points to continued cash conversion despite the stock selloff.

Amazon reported 39.3% growth at Amazon Web Services, but its third-quarter earnings release showed trailing 12-month free cash flow moving from a $14.8 billion inflow to a $42.4 billion outflow. The resulting $57.2 billion deterioration came despite accelerating cloud growth.

Microsoft is still converting cloud growth into cash, while Amazon’s infrastructure spending has outpaced its cash generation. Both companies are growing, but shareholders face different financial risks. Nvidia’s expanding financial commitments to AI customers add another uncertainty: how much industry demand depends on continued investment and financing arrangements.

What would settle the argument?

Burry may be enjoying a victory lap, but his warnings have not established that the entire industry is heading for collapse. He deserves credit for identifying vulnerabilities in the AI trade before the latest selloff.

Investors should distinguish companies that earn cash from AI demand from those that borrow heavily to build capacity. Profitable AI businesses with positive free cash flow, manageable debt, and customers capable of funding continued growth have the clearer financial footing.

The next test comes when the industry’s massive infrastructure investments must start paying for themselves. If GPU rental prices weaken while borrowing costs and depreciation rise, cash generation will determine which operators can fund the next capacity build rather than merely finance the last one.

247wallst.com (Source)

Filed under

Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,201 articles

Read also

« Previous article