Michael Burry has spent the better part of this year picking apart Big Tech’s AI spending, one balance sheet at a time.
His newest warning goes further than any of his previous posts, arguing the entire industry is now repeating a pattern that has wrecked fortunes before.
The five companies at the center of his latest argument are Amazon, Meta, Alphabet, Microsoft and Oracle. Burry says the hundreds of billions they are spending will not show up as a problem for years. But when it does, it could hit hard.
Burry’s latest warning hits on Big Tech’s AI spending
In a Substack post published on September 24, Burry compared the current AI spending boom to past capital cycle bubbles, including the dot-com era. He pointed to net capital investment across S&P 500 companies reaching 2.07% of GDP. That is the highest level in roughly four decades outside the aftermath of the March 2000 Nasdaq peak.
Burry does not think the fallout from that spending shows up right away. He wrote that write-offs could emerge around 2028 or 2029, when he believes companies could face significant write-offs tied to the infrastructure they are building today, Stocktwits reported.
More Wall Street:
- Wall Street’s AI trade faces its biggest valuation test
- The next Wall Street shift is already underway
- Wall Street sends strong 4-word verdict on the stock market
Alphabet drew the sharpest attention in the post. Burry estimates the company is carrying nearly $900 billion in off-balance-sheet commitments and exposures, including purchase obligations and other financing-related exposures tied to its AI infrastructure buildout.
Meta was not far behind. Burry pegged the company’s uncommenced leases and purchase obligations at roughly $700 billion, with that figure potentially approaching $1 trillion as its broader future commitments are taken into account, Stocktwits reported.
The numbers behind the $3 trillion figure
Add it all up across Microsoft, Amazon, Alphabet, Meta and Oracle. Purchase commitments. Future leases. Guarantees backing third-party debt. Construction in progress. Various special purpose vehicles. Burry gets to roughly $3 trillion.
That spending shows no sign of slowing. Hyperscalers are on track to spend roughly $800 billion on capital expenditures this year. Combined annual spending is expected to cross $1 trillion in the year ahead. That pace keeps expanding the very commitments Burry is warning about, a dynamic he has described as leaving these liabilities in “hypergrowth mode,” according to TheStreet.
This is not the first time Burry has raised concerns about how that spending gets accounted for. He has argued that hyperscalers are extending the useful life of Nvidia chips well beyond their realistic two to three year replacement cycle. He estimates that move could understate industry-wide depreciation by roughly $176 billion between 2026 and 2028, according to TheStreet.
The earnings impact of that accounting choice, in Burry’s view, would not be evenly spread. He has estimated Oracle’s earnings could end up overstated by roughly 26.9% and Meta’s by about 20.8% by 2028, if the depreciation assumptions behind current reporting hold up.
Bloomberg / Getty Images
How Big Tech and Wall Street are responding
Nvidia sent a seven-page memo to Wall Street analysts and named Burry directly. Four to six years is the real GPU depreciation window, it said. The 2020 A100 is still running. Still getting paid for. Two to three years, Nvidia argued, is not the cycle.
Nvidia also disputed one of Burry’s specific figures, stating it has repurchased $91 billion worth of shares since 2018 rather than the $112.5 billion Burry cited, arguing he had incorrectly included taxes on restricted stock units in his calculation, TheStreet reported.
Micron’s chief business officer told investors on a recent earnings call that customer demand for memory chips remains well above the company’s ability to supply across nearly every product category through 2028. That runs directly counter to the idea that current AI demand reflects speculation rather than genuine need.
Burry has continued building short positions that reflect his broader thesis, including bets against Oracle, Nebius, Micron and Palantir. Most of those stocks posted gains through August. That creates a clear test of his timing as the AI trade keeps rallying.
What it means for investors
Burry’s argument rests on a specific historical pattern. He has pointed to the late 1990s technology, media and telecommunications boom. After that cycle ended, depreciation and write-downs pushed aggregate S&P 500 net investment below zero for 12 consecutive quarters between mid-2003 and mid-2006.
That history is why Burry considers 2027 through 2029 the critical window. That is when he expects the accounting to catch up with the spending. It also underpins his leveraged short on the Philadelphia Semiconductor Index through January 2027 put options.
The disagreement between Burry and the companies he is targeting is unlikely to resolve quickly. Nvidia’s public rebuttal, Micron’s supply-demand data and continued capacity expansion from chipmakers like Taiwan Semiconductor all suggest the market still sees real demand behind the spending.
Burry disagrees. He keeps building the case that a reckoning is coming once the accounting catches up with reality.
Related: Michael Burry’s blistering message on economic inequality
