US Tech Giants' Hidden AI Debt Rose 90% to $3.2 Trillion in Six Months: What Nikkei Reported and Why Commenters Dispute the Number

US tech giants' off-balance-sheet obligations tied to artificial intelligence swelled roughly 90% in six months, reaching about $3.2 trillion — a figure equal to roughly 80% of those companies' assets, according to Nikkei Asia. The report, published November 4, 2026 (JST) and written by Kohei Yamada, frames the growth as evidence that investors are growing alarmed by seemingly unending AI spending.

The story reached the front page of Hacker News, where it drew 311 points and 135 comments. The discussion split sharply between readers who treat the $3.2 trillion as a systemic warning and readers who argue the number is a misleading aggregation of lease and power-purchase commitments that are neither debt nor due any time soon.

What the Nikkei Asia report actually says

Nikkei Asia's central claim is that off-balance-sheet debt from AI-related obligations at US tech giants almost doubled in six months to about $3.2 trillion, with the subheading stating that these commitments and obligations reach 80% of assets. The article's dateline is Palo Alto, California, and it names Google and Meta among the companies involved.

The full article body is behind a paywall and was not available in the scraped source. Only the headline, subheading, and opening paragraph are verifiable from the material at hand, so the specific accounting methodology behind the $3.2 trillion figure — which obligations are counted, over what horizon, and at what discount rate — is not described in the available text. A related Nikkei Asia article listed alongside it, titled "Five US tech giants' hidden debts soar to $1.65tn on opaque AI funding," indicates the same metric stood at roughly $1.65 trillion at an earlier point, which is consistent with the reported near-doubling.

The core dispute: is this debt at all?

The most-repeated objection in the Hacker News thread is definitional. Several commenters argue that signed contracts for future spending are not debt, and that lumping them together with borrowings overstates the risk.

"THIS IS NOT DEBT. With debt you are given money that you can spend, and then you have to pay interest for a while and then return all the money you spent. These are signed contracts for future spending. However, we don't know what the terms are for those contracts and how easy it is to get out of them." — HN commenter freediddy

The same commenter offered a concrete example of why contract terms matter: long-term compute contracts signed by Anthropic and Google with xAI's compute could reportedly be exited with three months' notice, which would make them a poor proxy for committed future spend.

A second objection concerns the time horizon. One commenter argued the figure bundles data center lease commitments and power payments extending decades into the future, and that even a total AI bust would not render the underlying data centers worthless, since non-AI server hosting still needs them.

"3.2 trillion sounds like a lot until you spread it over 20 years and half a dozen trillion+ dollar companies. It's hard to tell how much a bag they're holding but it's almost certainly negligible compared to revenue and market cap. The only real systemic risk is the risk of for some reason not needing any data centers." — HN commenter treis

The bull case: backlogs and remaining performance obligations

The strongest quantitative counterargument in the thread points to cloud backlogs. One commenter cited aggregate remaining performance obligations (RPOs) — revenue customers have committed to spend over the next three to six years — of $2.33 trillion across the tech giants, up from $0.5 trillion in 2022, with year-over-year growth ranging from 84% at Microsoft to 390% at Google. The same commenter said Microsoft and Google alone are expected to recognize roughly $0.5 trillion of that revenue within 12 to 24 months.

That commenter also acknowledged the concentration risk: OpenAI and Anthropic may account for around $1 trillion of those RPOs, leaving roughly $1.3 trillion of other demand. The conclusion drawn was that service providers generally try to stay ahead of the demand curve rather than behind it, and that Amazon, Microsoft, and Alphabet have long track records of fiscal discipline.

The bear case: cash flow, rates, and junk-rated newcomers

The pessimistic reading focuses less on the accounting label and more on funding capacity. One commenter argued the problem is not the commitments themselves but that the companies no longer have the free cash flow to cover them, and are therefore raising cash in what is described as a higher interest rate environment.

"For the new players like Anthropic and their likely 'junk' credit rating this is going to be incredibly painful. It's basically like playing in a casino on margin. If it works, great. If not you are totally f!*ked." — HN commenter cmiles8

Another commenter framed the same concern in macroeconomic terms, arguing that the market's liquidity should not be stretched to fund a handful of companies and that failing firms should be allowed to fail. A separate commenter predicted bailouts would follow.

Scale comparisons offered in the thread

Commenters reached for yardsticks to make $3.2 trillion legible. One noted that total US government debt is about $40 trillion. Another noted that total 2025 US income-tax revenue was approximately $5.3 trillion. A third observed that the figure is large enough to become a recurring talking point for macro commentators.

The counter-counterargument: maybe the CFOs know something

Several commenters pushed back on the premise that the spending is irrational. One pointed out that the chief financial officers and chief executives making these commitments have access to the best consumer-side metrics and internal visibility into where the industry will be in one to two years, and may simply be betting that AI will absorb knowledge work the way software once did, only faster.

"When I worked at Apple, I remember seeing the commentariat insisting that was bound to fail, that we didn't know what we were doing. When in fact extremely competent people knew precisely what they were doing." — HN commenter xvector

Another commenter described the situation as a huge bet on AGI in which the winner reaps enormous wealth and power, while noting skepticism that the underlying technology can support it yet. A third asked the practical question underlying the whole debate: if the big labs fail, can the excess data center capacity be redeployed profitably for inference?

What the debate does not resolve

Three questions remain open after reading both the report and the thread. First, what exactly is inside the $3.2 trillion — the available article text does not itemize which obligations are counted or over what period. Second, how cancellable are those obligations, since contract terms determine whether they behave like debt or like options. Third, how much of the committed cloud backlog is concentrated in a small number of AI labs whose own creditworthiness is unproven.

Until those are answered, the $3.2 trillion figure is best read as a measure of announced intent and contractual exposure rather than as a balance-sheet liability — a distinction that the Hacker News discussion makes repeatedly and that any citation of the number should preserve.