NEW YORK, July 21 — A new analysis has put a number on something many in finance have long suspected: the true debt of America’s biggest tech companies is far larger than what appears on their balance sheets. The figure is staggering—$1.65 trillion in hidden borrowing, much of it tied to the race to build artificial intelligence systems. But behind the headline is a story about financial engineering that is legal, widely used, and only now beginning to attract serious scrutiny.
The technique goes by the dry name of off-balance-sheet financing. In plain terms, it means companies find ways to fund massive projects—like AI data centers—without recording the associated debt as a liability.
The result, according to the analysis that aggregated instruments across five major US tech firms, is that these companies are far more leveraged than their official financial statements suggest.
The Hidden Debt
To understand the scale, consider this: the same five companies reported roughly $200 billion in long-term debt in their 2023 annual reports. The $1.65 trillion in off-balance-sheet obligations is more than eight times that amount. That gap is large enough to mislead credit rating agencies and distort bond pricing, the analysis notes.
The most common method is the operating lease. When a tech company rents data-center equipment, it treats the payments as ongoing operating expenses rather than as debt.
Under current GAAP rules, that is perfectly permissible—but it hides the true commitment. Another approach involves special purpose vehicles, or SPVs, which isolate AI investments legally from the parent company’s balance sheet. That keeps the debt off the books even when the parent is effectively on the hook.
Then there are convertible notes and structured equity deals with AI startups. The debt is not recorded until and unless it converts into equity.
Microsoft’s investment in OpenAI, for example, involved a multi-billion-dollar convertible note that does not appear as debt on Microsoft’s balance sheet. Amazon’s investment in Anthropic took a different form: a commitment to purchase cloud credits, booked as a future expense rather than a liability.
Legal but Opaque
None of these practices are illegal. The Financial Accounting Standards Board has debated tightening the rules, but no changes have been finalized. The result is a system where shareholders and regulators are left to guess at the true risk profile of some of the world’s most valuable companies.
For investors, the stakes are personal. A company that looks conservatively financed might actually be carrying enormous hidden obligations.
Bondholders, in particular, could be blindsided if a sudden reassessment of off-balance-sheet debt leads to credit downgrades or higher borrowing costs. And for regulators, the question is whether the current accounting standards are doing their job—or whether they are allowing a multitrillion-dollar blind spot to persist.
What Comes Next
The FASB debate over tighter rules has not produced a resolution. Until it does, the gap between reported debt and real debt will remain. And as the AI buildout accelerates, that hidden leverage is only likely to grow.
For now, the numbers are out in the open—at least for those willing to look behind the balance sheet.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Tech Giants Hide AI Debt via Off-Balance-Sheet Tricks”, “description”: “NEW YORK, July 21 — A new analysis has put a number on something many in finance have long suspected: the true debt of America’s…”, “datePublished”: “2026-07-21T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “James Roberto”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.2bme.xyz/”}}
