US tech giants' off-balance-sheet AI obligations nearly double to $3.2 trillion in six months

Off-balance-sheet obligations tied to artificial intelligence at major US technology companies have nearly doubled in about six months, swelling to roughly $3.2 trillion, according to a Nikkei Asia report published on November 4, 2026. The commitments — spanning Google, Meta and other large operators — now equal about 80% of those companies' assets, and investors are growing increasingly uneasy about spending that shows no sign of letting up.

The scale of the obligations

Nikkei Asia puts the total of AI-related commitments and obligations at roughly $3.2 trillion, up around 90% in roughly half a year. These are not conventional loans: they are contractual promises to pay in the future, which is why the report frames them as a form of hidden debt. Measured against the companies' assets, the obligations come to about 80%, according to the outlet.

The buildout behind the numbers is physical as well as financial. The report points to continued construction of AI infrastructure across the United States — Nikkei illustrates the story with an Amazon data center in Indiana — as technology giants pour capital into the computing capacity needed to train and run AI models.

Why the obligations are off the balance sheet

Off-balance-sheet commitments do not show up as debt on the face of a balance sheet. They typically appear in the notes to financial statements as purchase commitments and similar contractual obligations. That means headline leverage metrics can look modest even when a large share of future cash flow is already spoken for. A near-doubling in six months, as Nikkei reports, is a change that most top-line debt figures would not capture.

Investors are watching

According to Nikkei, unease among investors is mounting as the spending spree continues without an obvious end point. The report drew quick attention in technical circles, reaching the front page of Hacker News — a rough proxy for how closely the industry is tracking the economics of AI infrastructure.

Why it matters

A $3.2 trillion pile of commitments approaching 80% of assets is a striking gauge of how aggressively US tech giants are locking themselves into future AI spending. Three implications stand out.

First, transparency: because these obligations live outside headline debt, anyone relying on conventional leverage ratios is seeing an incomplete picture. The gap between reported debt and total commitments is now wide enough to change how the sector's financial health reads.

Second, rigidity: contractual commitments must be honored whether or not AI revenue grows as hoped. If demand falls short of expectations, the payouts still come due, and that could squeeze room for other priorities — hiring, research, buybacks — or weigh on credit profiles.

Third, momentum: a 90% increase in six months is itself a signal. Whatever the eventual payoff of AI, the commitments suggest companies are binding themselves to capacity at a pace that outstrips what their balance sheets alone would show — and investor patience for that trajectory is, by Nikkei's account, wearing thin.