Michael Burry is not the type to miss an opportunity to say “I told you so.” And right now, he has a good reason. A Wall Street Journal investigation found that the four biggest names in AI, Alphabet, Amazon, Meta, and Microsoft, have quietly accumulated $3 trillion in future AI-related financial obligations that don’t appear on their official balance sheets. Burry, best known for predicting the 2008 mortgage collapse, responded on X with a pointed reminder that he had been flagging this exact issue months earlier, in 2025.
The numbers the WSJ published are worth sitting with. The four companies report a combined $248 billion in lease liabilities and $356 billion in long-term debt through standard disclosures. Those figures sound large. But underneath them sits roughly $1.2 trillion in lease commitments that haven’t yet started, plus another $1.9 trillion in purchase commitments. That’s the actual scale of what’s been quietly locked in while the public narrative focused on quarterly capex announcements and data center expansion press releases.
Alphabet is the most exposed of the group by a significant margin. Its off-balance-sheet purchase commitments alone total $811 billion. Meta follows at $349.3 billion, Microsoft at $228.6 billion, and Amazon at $130.1 billion. These are not speculative projections. These are signed financial obligations, money that has been committed to suppliers, chip manufacturers, and infrastructure partners, just not reflected in the line items most analysts and investors typically focus on.
Burry’s post on X didn’t stop at taking credit. He also pointed to what he called “compression’s threat” as the next risk that mainstream financial media will only catch up to in 2027. He didn’t elaborate publicly, but the implication is clear enough. When you front-load trillions in capital into AI infrastructure before the revenue models are fully proven, you create the conditions for a painful correction when growth assumptions get revised downward. That’s not a fringe view. It’s a concern several infrastructure analysts have raised quietly while the public conversation stayed fixated on model releases and benchmark scores.
For developers and founders, the practical takeaway is less about stock movements and more about what this spending structure signals for the cloud and AI services market. When Alphabet, Microsoft, and Amazon have committed this much capital to AI infrastructure, they are not going to let those assets sit idle. That means continued pressure to push AI consumption through their platforms, aggressive pricing to drive adoption, and product decisions shaped by the need to justify those commitments. The incentives are baked in now. So the services landscape these companies offer over the next three to five years will be shaped at least partly by the need to monetize $3 trillion in locked-in obligations. That context matters when evaluating which platforms to build on and which vendor relationships carry real staying power.
