Michael Burry Flags $3 Trillion Risk in Big Tech’s AI Infrastructure Binge

Key Takeaways

• The investor famous for ‘The Big Short’ claims Amazon, Meta, Alphabet, Microsoft and Oracle hold approximately $3 trillion in AI infrastructure obligations.
• Burry draws parallels between today’s AI investment surge and the dot-com crash, predicting significant write-downs by 2028-2029.
• In response to Burry’s assertions, Nvidia issued a detailed seven-page rebuttal challenging his chip depreciation calculations.
• The hedge fund manager has expanded short bets on Micron, Nebius, Palantir and various semiconductor companies.
• Simultaneously, Burry is accumulating value positions in beaten-down names including Build-A-Bear, Birkenstock and Sprouts Farmers Market.

Michael Burry has delivered a stark new warning regarding the massive capital outlays tech giants are pouring into artificial intelligence infrastructure. The legendary investor, who famously anticipated the 2008 financial crisis, believes today’s spending trajectory mirrors historical bubbles that culminated in significant losses.

The hedge fund manager outlined his concerns in a detailed Substack analysis released on September 24. His examination centered on five major technology corporations: Amazon, Meta, Alphabet, Microsoft and Oracle.

The $3 Trillion Question

According to Burry’s analysis, these tech behemoths collectively shoulder nearly $3 trillion in financial commitments related to AI infrastructure development. These obligations encompass purchase contracts, future lease arrangements, financial guarantees and ongoing construction expenses.

In his examination, Alphabet emerged as the company with the largest exposure. Burry’s calculations suggest the search giant maintains approximately $900 billion in off-balance-sheet obligations connected to its AI expansion plans.

For Meta, Burry tallies uncommitted leases and purchase commitments at around $700 billion. He notes this figure could balloon toward $1 trillion when accounting for the full scope of contractual obligations.

Drawing historical comparisons, Burry highlighted how net capital expenditures across S&P 500 constituents have reached levels as a percentage of GDP not seen in roughly forty years—reminiscent of the late-1990s internet boom.

The investor doesn’t anticipate immediate consequences. Rather, he forecasts that material asset impairments will surface sometime around 2028 or 2029.

Beyond the sheer scale of spending, Burry questions how tech companies are depreciating their Nvidia GPU investments. He contends that cloud providers are extending the useful economic life of these processors well beyond a realistic two-to-three-year utilization period.

This accounting methodology, according to his analysis, potentially conceals approximately $176 billion in depreciation expense across the sector between 2026 and 2028.

Tech Companies Fire Back

Nvidia didn’t remain silent on Burry’s assertions. The chipmaker distributed a comprehensive seven-page document to Wall Street research analysts, maintaining that a four-to-six-year depreciation schedule more accurately reflects the operational lifespan of its processors.

The company also corrected what it characterized as a factual error, noting its actual share buyback total since 2018 amounts to $91 billion, significantly less than Burry’s stated $112.5 billion figure.

Meanwhile, Micron’s chief business officer publicly stated that memory chip demand continues to exceed the company’s production capacity through at least 2028. This outlook directly contradicts Burry’s position that current AI-driven demand is artificially inflated.

Burry has backed his analysis with capital. His disclosed short positions include Oracle, Nebius, Micron and Palantir.

Additionally, he’s established short exposure to the Philadelphia Semiconductor Index via put options expiring in January 2027. Notably, the majority of his short targets have delivered positive returns through August.

Burry isn’t alone in his skepticism. GMO’s Jeremy Grantham has publicly characterized AI valuations as exhibiting bubble characteristics. Similarly, DoubleLine Capital’s Jeffrey Gundlach has predicted that clear winners and losers will eventually emerge from the AI investment surge.

Value Opportunities in Burry’s Portfolio

Despite his bearish stance on AI infrastructure, Burry isn’t entirely defensive. He’s initiated complete positions in five companies operating outside the AI ecosystem.

His recent purchases include roofing materials distributor QXO, natural foods retailer Sprouts Farmers Market, specialty toy company Build-A-Bear, footwear manufacturer Birkenstock and Latin American e-commerce platform Mercado Libre.

Each of these positions represents stocks that have declined significantly year-to-date. Build-A-Bear has suffered the steepest drop, plummeting as much as 57%.

Micron’s upcoming quarterly results may prove pivotal. Analyst consensus anticipates year-over-year earnings growth approaching 940%, per Investor’s Business Daily data. The company’s performance and forward guidance could substantially influence market sentiment regarding the sustainability of AI-related capital spending.