Michael Burry warned that the AI capex boom could collapse like the 2008 housing market, citing systemic overinvestment and circular financing patterns.

Investor Michael Burry, best known for predicting the 2008 financial crisis, is warning of fresh risks in the rapid expansion of AI infrastructure spending across major U.S. tech companies.

Burry highlighted a post from Apollo Global Management partner and Chief Economist Torsten Slok, calling it "three great charts" that draw a striking comparison between the current AI data center buildout and the housing boom-and-bust of the 2000s.

Slok projects that cloud companies' capital expenditure will reach roughly 3% of U.S. GDP annually from 2027 to 2029—a sharp rise from 0.3% of GDP in 2019 and 1.4% in 2025. This projected level is twice the peak of the late-1990s telecom and fiber buildout, which topped out at 1.2% of GDP in 2000 before collapsing and dragging the economy into recession.

Data center investments remain less than half the size of residential investments at their 2005 peak of 6.6% of GDP, but the pace of increase is far faster than either previous cycle.

"The bottom line is that the data-center buildout is smaller than housing in level but larger in the change in share of GDP, and faster than either previous cycle," Slok wrote. He noted that housing's unwind from 6.2% of GDP in early 2006 to 3.0% by the end of 2008 was what made that recession so severe, while telecom's smaller reversal produced only a mild downturn.

"A cycle that builds at 0.85 percentage points a year can unwind at a similar pace, and that, rather than the buildout itself, is the macro risk if AI demand disappoints," Slok cautioned.

Five of the largest U.S. tech companies—Amazon, Alphabet, Meta Platforms, Microsoft, and Oracle—are expected to spend approximately $800 billion in combined capital expenditures this year, with the majority directed toward new data center construction.

S&P said in May that U.S. hyperscalers' accelerating AI investments could reach $1 trillion in 2027, driven by AI demand, rising component costs, and capacity expansion, with Microsoft, Alphabet, Amazon, and Meta all continuing to lift spending. Bank of America raised its estimates following Big Tech earnings and now expects aggregate hyperscaler capex to exceed $1 trillion in 2027, with 2026 spending estimated at $800 to $900 billion—a view echoed by Evercore ISI, Raymond James, and Moody's.

Peter Diamandis, executive chairman of the X Prize Foundation and an early Tesla investor, stated on X that 2,441 proposed data center projects in the U.S. are scheduled to break ground between 2026 and 2028, representing as much as $2.48 trillion in planned investment. "That pipeline is roughly 5 times the cost of building the original Interstate Highway System (inflation-adjusted)—except this infrastructure doesn't move cars; it moves intelligence," Diamandis said, without providing a source for his data.

Burry has positioned his portfolio to reflect his skepticism, building or expanding bearish positions in Nvidia, Palantir, Caterpillar, Applied Materials, Micron, Oracle, and Nebius since late 2025. He has also rolled some Nvidia puts into 2027, reduced his Palantir put exposure while maintaining a direct short, and exited his Microsoft long position, signaling deepening conviction that AI infrastructure valuations remain dangerously stretched.

Outside the AI trade, Burry more than doubled his stake in Flutter Entertainment after a post-earnings selloff, and maintained or added to positions in Fiserv, Zoetis, and Mercado Libre, which he views as attractive long-term value opportunities despite near-term headwinds.