When investors look at the balance sheets of Alphabet, Microsoft, Amazon, Meta, and Oracle, they see debt that appears large but manageable. However, a deeper examination reveals a second, much larger pile of debt lurking off the books. According to a Nikkei study, this hidden figure has reached $1.65 trillion, nearly eight times higher than four years ago and surpassing the $1.35 trillion the five companies report outright. This debt is tied up in off-balance-sheet vehicles, the same kind of structure that Enron infamously used to hide its liabilities before collapsing 25 years ago. While Enron's actions were fraudulent, today's tighter regulations and fuller disclosures make these practices legal—though not without controversy.
The Enron echo and its implications
The comparison to Enron is striking. Enron used special purpose vehicles (SPVs) to conceal debt from its balance sheet, leading to one of the largest corporate bankruptcies in history when the scheme unraveled. Today, tech giants employ similar structures to finance their massive AI infrastructure investments without recording the associated liabilities directly. Accounting expert Gil Luria told Bloomberg Law, "Enron's crime wasn't having special purpose vehicles. Enron's crime was hiding them." The key difference now is that companies are required to disclose these off-balance-sheet arrangements in footnotes, making them technically transparent. Yet, the sheer scale of the hidden debt—more than the combined market capitalization of many countries—raises red flags among analysts and regulators.
How the off-balance-sheet mechanism works
The mechanics are straightforward but powerful. A company packages its debt for chips, servers, and power into a separate legal entity, often a joint venture with a financial partner. By doing so, the cost never flows through the company's own accounts. For example, Meta's Hyperion data center in Louisiana is financed through a structure where Meta and Blue Owl Capital both contribute equity. The entity then takes on $27 billion in debt. Meta is the sole tenant but argues it does not need to record that debt because it is not responsible for finding replacement tenants if the data center becomes vacant. Oracle has $260 billion in future lease commitments that will eventually land on its books. Nvidia carries $119 billion in purchase obligations for chips and equipment. Alphabet and Microsoft also keep their vehicles off-book, though they provide some disclosure in annual reports.
The scale of the off-balance-sheet debt
The numbers are staggering. Meta's off-balance-sheet debt alone is about $420 billion, nearly three times its reported debt. Oracle's has grown roughly thirtyfold in four years, reaching $260 billion. Amazon's hidden debt is estimated at $340 billion, Microsoft's at $280 billion, and Alphabet's at $150 billion. These figures represent the total future obligations for data centers, servers, and energy contracts. The industry as a whole is expected to spend more than $3 trillion through 2028 building and equipping AI data centers, much of it financed against the chips inside them. This race to dominate AI has led to a debt boom that rivals the subprime mortgage crisis in both complexity and potential risk.
Why it matters now
The timing is particularly awkward. Four of the five companies report earnings in the next two weeks, and their reported debt will look tidy and manageable. The $1.65 trillion sitting in the footnotes will not make the headlines. The catch is what happens later. When a data center goes live, its lease rolls onto the balance sheet at once. If AI demand falls short, the facility is marked down, and the loss lands on the lenders and insurers who funded it. Some already see the risk. S&P Global has cut Oracle's credit rating over stretched leverage, and both Morgan Stanley and Moody's have flagged the wider issue. "What if one of these companies was a house of cards," asked accounting consultant Tom Selling, "and was propping itself up with this accounting treatment?" The question is especially pertinent given the speculative nature of AI investments and the possibility of a bubble bursting.
Historical context and regulatory evolution
The use of off-balance-sheet financing is not new. Enron's collapse led to the Sarbanes-Oxley Act and tighter rules on SPVs. However, the Financial Accounting Standards Board (FASB) has since allowed certain exemptions for operating leases and joint ventures. The current rules require companies to disclose off-balance-sheet arrangements in footnotes, but they do not require them to be consolidated on the balance sheet if the company does not control the entity. This loophole has been exploited by tech giants to keep their debt ratios low, making them appear less leveraged to investors and credit rating agencies. Critics argue that this practice understates risk and could lead to a systemic shock if a major player defaults.
Detailed analysis of each company
Meta (formerly Facebook) has been particularly aggressive in using off-balance-sheet structures. Its data center investments in Louisiana and elsewhere are financed through joint ventures with infrastructure funds like Blue Owl Capital. Meta's total off-balance-sheet debt is estimated at $420 billion, compared to its reported debt of $150 billion. The company argues that since it is only a tenant, the debt should not appear on its books. Oracle's $260 billion in off-balance-sheet lease commitments are even more dramatic. The company has grown its cloud infrastructure rapidly, but its leverage has raised concerns. Amazon's hidden debt stems from its massive data center build-out for AWS, while Microsoft and Alphabet use similar structures for their Azure and Google Cloud investments. Nvidia, while not a cloud provider, carries $119 billion in purchase obligations for chips, some of which are financed through off-balance-sheet vehicles.
Potential risks and investor implications
Investors reading this week's earnings reports will see less than half the true leverage of these companies. The hidden debt is not immediately due, but it represents future cash outflows that could strain finances if AI adoption slows. The data center boom is supported by assumptions that AI demand will continue to grow exponentially. However, if AI fails to deliver on its promises or if a more efficient technology emerges, the value of these assets could plummet. The lenders and insurers who financed the structures would bear the initial losses, but they could pass them on to the tech companies through higher costs or litigation. Moreover, the off-balance-sheet debt could suddenly appear on the balance sheet in a crisis, triggering rating downgrades and stock sell-offs.
Industry watchers point to parallels with the dot-com bubble, where companies spent heavily on infrastructure that later became obsolete. The difference this time is the scale of debt and the concentration in a few players. The five tech giants now account for a significant portion of global data center capacity, and any failure could have ripple effects across the economy. The use of off-balance-sheet vehicles also makes it harder for regulators to gauge systemic risk. The Financial Stability Oversight Council has started to examine the issue, but no action has been taken yet.
The honest version and what companies say
None of this is illegal, and the companies insist future earnings will more than cover the bills. The disclosures exist, buried in footnotes of annual reports, for anyone willing to dig. Meta, for example, states in its 10-K that it has "off-balance-sheet arrangements" related to data centers, but provides only limited detail. Oracle's and Microsoft's disclosures are similarly opaque. The companies argue that these structures are standard practice for large infrastructure projects and that they are better off taking on debt through separate entities to isolate risk. However, as the hidden debt grows, so does the potential for a crisis of confidence. In a year of loud bubble talk, the quiet number hiding off the page may be the most important indicator of all.
The coming weeks will see earnings reports that highlight strong revenue growth and managed debt levels. But the $1.65 trillion in hidden obligations will remain in the footnotes, waiting for the next downturn to reveal the true financial health of the AI industry.
Source:TNW | Amazon News