Big Tech's AI commitments reach $3tn, about five times its annual capital spending

Most of it is signed leases and purchase contracts in financial-statement footnotes, and not all of it need ever be paid.

The figure comes from a Wall Street Journal analysis, relayed by Toobit and the AI Fire newsletter, covering nine companies: Alphabet, Meta, Microsoft, Amazon, Oracle, Nvidia, Broadcom, SpaceX and AMD.

The mix matters. By the same relayed analysis, about $1.2tn is signed leases that have not yet started and about $1.9tn is purchase commitments. Alphabet alone disclosed about $811bn of purchase commitments and contractual obligations as of June 2026, and Meta about $347bn of leases not yet started. Under current accounting rules these sit in footnotes, not in reported debt.

Against that, conventional capital spending looks modest. The analysis, as AI Fire relays it, set the $3tn beside about $600bn of capex over the latest 12 months, roughly five times as much. But a commitment is not a payment. The total is a ceiling on what has been pledged, not a measure of cash out the door, and Axios noted that these are often contracts that could in theory be renegotiated, with no clear period over which they fall due.

Other estimates land in the same region. Morgan Stanley, as Axios reported on August 27, put off-balance-sheet commitments of seven companies at about $3tn, on top of an estimated $770bn of debt and lease obligations already on balance sheets. An October 22 FourWeekMBA analysis, reported by Tech Startups, reached about $3.56tn, including $454.52bn of noncancelable commitments at Meta in its September-quarter filing.

The debate has therefore moved from whether these companies will keep spending on AI to how much of the spending is already locked in. The test comes over the next few quarters: how much of the total appears in reported capital spending, and whether companies trim these contracts if AI demand disappoints.