Tom Lee dismisses Burry's Enron warning on $3 trillion AI trade

Fundstrat's Tom Lee pushed back on Michael Burry's Enron-style warning over the AI trade, arguing that the $3 trillion in off-balance-sheet commitments disclosed by nine technology giants misrepresents the real risk to investors.

"The $3 trillion in off-balance-sheet deals scaring Wall Street tells investors little about the real risk," Lee, head of research at Fundstrat Global Advisors, said.

The Wall Street Journal analysis that triggered the exchange found Alphabet, Amazon, Meta and Microsoft report $248 billion in lease liabilities and $356 billion in long-term debt on their balance sheets, but hold an additional $1.2 trillion in uncommenced leases and $1.9 trillion in purchase commitments covering chips, energy and data-center infrastructure. Alphabet alone accounts for $811 billion in hidden purchase commitments, while Meta holds $349.3 billion, Microsoft $228.6 billion and Amazon $130.1 billion.

The clash between two of Wall Street's most-watched voices on AI valuations comes as the sector's biggest spenders face intensifying scrutiny over whether the infrastructure buildout can generate returns before financing costs rise. Burry has warned of a three-way "compression" — demand weakening, earnings falling as expenses catch up, and financing tightening — that he expects to unfold as the off-balance-sheet obligations come due.

Burry's Long-Running AI Bear Case

Burry, best known for correctly predicting the 2008 housing crisis, first flagged the AI spending risk in late 2025 after returning to X following a two-year hiatus. He alleged that hyperscalers were understating depreciation by assigning five- to six-year useful lives to Nvidia chips and servers whose economic cycle, in his view, is closer to two or three years. He estimated the practice could suppress depreciation expenses by $176 billion between 2026 and 2028, inflating Oracle's 2028 earnings by 26.9 percent and Meta's by 20.8 percent.

In May, Burry published a Substack post titled "Tracepalooza & the Bezzle," arguing that training and benchmarking activity could exaggerate near-term AI demand and create a bullwhip effect. He highlighted Nvidia's $182 billion in forward purchase commitments, offshore financing, and construction assets that do not begin depreciating until placed in service.

Burry has since disclosed short positions in Oracle at around $145 per share and a larger short in Nebius at $212 per share. He continues to hold shorts in the iShares Semiconductor ETF, Micron, Nvidia, Caterpillar, Palantir, Tesla and Applied Materials.

What the $3 Trillion Actually Covers

The nine companies — Alphabet, Amazon, Meta, Microsoft, Oracle, Nvidia, Broadcom, SpaceX and AMD — disclosed the commitments in filing footnotes rather than on the face of their balance sheets. The $1.9 trillion in purchase commitments covers chips, energy and data-center infrastructure, while the $1.2 trillion in uncommenced leases represents facilities contracted but not yet occupied.

These obligations compare with roughly $600 billion in capital expenditures across the companies' latest reported 12-month periods, meaning the forward commitments represent roughly five years of current spending levels.

For investors, the debate carries direct implications for AI-exposed equities. Alphabet shares have gained 9.48 percent year to date, while Microsoft is up 2.44 percent. Meta has fallen 10.64 percent and Amazon has risen 13.79 percent. Burry's warnings have not yet triggered a broad selloff, but the divergence between the two camps leaves the AI trade exposed to sharp repricing if either side proves correct.

This article is for informational purposes only and does not constitute investment advice.