AI Infrastructure Hidden Debt Reaches $1.65 Trillion as Credit Risks Mount for Big Tech

AI tech companies have 'hidden debt' worth around $1.65 trillion, report claims -- amount is 122% of debt reflected on the balance sheets of Alphabet, Amazon, Meta, Microsoft, and Oracle

Five U.S. tech giants heavily invested in AI and its related infrastructure reportedly have an estimated $1.65 trillion in hidden debt, with the figures annotated in their quarterly financial statements instead of being listed in their balance sheets. According to Nikkei Asia, this is higher than the $1.35 trillion officially listed, meaning investors could be caught unaware once the hidden figures come to light. The publication says that Meta has a high off-balance-sheet-to-recorded-debt ratio, with the company owing $420 billion in unlisted debts compared to the $140 billion written on the balance sheet. Oracle purportedly also has a massive $273.3 billion of hidden debt, which is a 2,900% jump from the hidden debt it had from 2022. This may sound strange, but it's actually an accepted accounting practice. The "hidden debt" stems from long-term contracts that have been signed but have not come into force yet, which, Nikkei says, is mostly related to the billions of dollars promised to data center operators. The AI race has got many hyperscalers signing contracts and agreements with data center operators, saying that they will pay for the compute they generate once their project goes online. While any institution promising to pay any amount of money for services or goods delivered is obliged to list them as a liability, the fact that these data centers haven't started operations means that these agreements are off-the-books at the moment. But when these projects come online, the contracts that the tech giants have signed will come into force, and they'll have to pay for the compute that these sites will deliver, no matter if there is demand or not. Nevertheless, these tech companies aren't just pouring money into future contracts just for the sake of it. Alphabet, Amazon, and Microsoft reportedly have a cloud service backlog worth $1.45 trillion, meaning these are services yet to be rendered and paid. Amazon Web Services CEO Matt Garman also told the publication that the investments that the company is getting into are "not speculative." While this may seem like a good way to secure capacity -- sign customer contracts that guarantee demand and then enter into long-term agreements with data centers to get the compute needed to deliver the services- it opens these tech giants to massive amounts of risk. That's because if the demand fails to materialize, then they'd be left paying for excess compute without having any customers to sell them to. What's more alarming is that Nikkei says that these investment expenditures are exceeding their earnings, meaning these big tech companies are increasingly relying on corporate bonds and new shares to fund them. Even though demand for AI compute is increasing, it's still a relatively new and unproven technology, with many experts saying that it should benefit more people to avoid a bubble. The cost of using AI for nearly everything, called "tokenmaxxing," has also caught some companies by surprise, with agentic AI eating up annual AI budgets in a matter of weeks. Because of this, some companies are reducing their use of AI or are switching to more affordable models from China. This uncertainty, paired with the way tech companies "hide" these liabilities, is quite concerning, as they would appear to have less long-term obligations than they actually do. This isn't the first time that an industry giant has used similar accounting techniques. The publication cited Enron's 2001 collapse, which was due to the company hiding its troubled assets through special purpose entities and marking unrealized gains from trading contracts into its current income statements. While the tech giants are not hiding underperforming assets off their balance sheets and committing fraud, they're still using a similar mechanism to list their upcoming obligations. Although these are technically not debt, they still behave like one, and the way they're reported is what's concerning some experts. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

AI boom's hidden $500B debt look eerily similar to the 2008 mortgage crisis

AI's explosive growth sits a massive debt structure that may blow up in our faces * Bloomberg pegs outstanding AI data center debt above $500 billion right now * CoreWeave isolates each loan inside its own separate special purpose vehicle * Parent companies report only a fraction of their real total exposure, hiding the rest in shell entities A growing body of analysts now warns that AI data center debt increasingly resembles the subprime mortgages that triggered the 2008 financial crisis. Much of that debt is issued through special purpose vehicles, structures that keep billions of dollars off corporate balance sheets entirely. Bloomberg estimates more than $500 billion in outstanding AI data center debt, with roughly $200 billion held by private credit funds. A debt structure built on theoretical revenue Special Purpose Vehicles (SPVs) raise debt to build data centers, then repay creditors only once paying customers begin generating revenue. That structure is exactly why CoreWeave has raised billions through separate SPVs for individual loans, including an $8.5 billion facility tied to Meta's contract, since each loan stays isolated inside its own entity. The same logic explains why Nikkei Asia reported that Meta, Google, Amazon, Microsoft and Oracle have accrued around $1.65 trillion in debt over five years, much of it spread across similar vehicles rather than sitting on any single balance sheet. That gap between real exposure and reported debt exists because these vehicles are jointly owned with outside investors, letting the parent company report only a fraction of the risk. Meta's Hyperion data center shows the pattern clearly: it is owned 80% by Blue Owl and only 20% by Meta itself, so the bulk of the debt lives with Blue Owl on paper even though Meta is the intended tenant. Google has used the same approach, backstopping debt-funded data centers built by Fluidstack, Cipher Mining and TeraWulf without those obligations ever touching its own balance sheet. That kind of arrangement is precisely what drew scrutiny from auditor Ernst & Young, which flagged Meta's structure as a critical audit matter, questioning who ultimately bears its economic risk. The stakes extend well beyond the companies involved, because pension funds and insurers are also directly exposed, with many now relying on data center returns to fund future payouts. Echoes of the 2008 mortgage collapse The comparison to 2008 holds up because both bubbles rested on the same flawed premise: that demand would keep growing forever and never needed to be tested. Subprime mortgages were the proof of that thinking at the time, and by 2006 roughly 20% of all new mortgages issued in the United States were already classified as subprime, according to government data. Rather than treat that as a warning sign, financial institutions bundled those loans into complex securities, a move that obscured the true underlying risk from investors and rating agencies alike. Financier Michael Milken captured the mood of the era when he publicly described such securities as a "financial innovation" that would broadly increase national prosperity and jobs. Reality caught up with that optimism once mortgage defaults began rising sharply in 2005, and the damage cascaded through the entire financial system from there. Lehman Brothers embodied how unchecked that confidence had become, operating at more than 25 times leverage in 2005 without serious pushback from regulators or rating agencies. Today's numbers echo that same pattern of unexamined risk: analysts estimate more than $1.4 trillion in bank exposure to private credit, with $300 billion of it held by major banks alone. Some estimates suggest planned AI data center capacity exceeds actual annual compute demand by a factor of roughly 15 times. Unlike 2008, this risk is not driven by derivatives but by the sheer scale of individual data center construction costs. Whether this debt unwinds gradually or all at once likely depends on how quickly major AI customers can pay their bills. For now, the scale of exposure across banks, pensions and insurers suggests the comparison to 2008 is not merely rhetorical. Via Ed Zitron Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

AI Companies Are Trying to Hide a Staggering Amount of Debt

Can't-miss innovations from the bleeding edge of science and tech AI companies are pouring untold billions of dollars into enormous data centers in their efforts to sustain increasingly complex and resource-intensive AI models. It's an extremely costly undertaking built on seemingly bottomless hype -- and a mountain of debt. As Japanese financial newspaper Nikkei Asia found in a recent investigation, just five US tech giants -- Alphabet, Microsoft, Amazon, Meta, and Oracle -- are hiding an estimated $1.65 trillion in debt that doesn't appear on balance sheets. That's even more than the $1.35 trillion in debt the five companies officially reported in their financial data for the most recent quarter. Meta alone has amassed around $420 billion in off-balance-sheet debt, according to Nikkei, highlighting how precarious the AI industry's steep investment in AI has become, and inspiring comparisons to energy company Enron, which collapsed in spectacular fashion in 2001 because of similar debts hidden behind shell companies. Like Enron, they're using special purpose vehicles, or off-balance sheet arrangements such as legally distinct subsidiaries, as a way to make their financial reporting look healthier than it actually is -- often a glaring sign that something is deeply amiss behind the scenes. "The accounting treatment itself is in fashion," technical accounting consultant Tom Selling told Bloomberg. "But what if one of these companies was a house of cards and was propping itself up with this accounting treatment? To me, that's the risk." Experts continue to warn of an AI bubble, noting the enormous and widening gulf between company valuations and their comparatively measly profits. The latest news will do little to quiet critics who say the situation is more dire than the companies' official balance sheets suggest. To keep up with the ongoing AI race, tech giants are committing vast sums to build out large-scale data center projects, a long-term bet that may -- or may not -- pay off. They're also selling new shares to raise new funds, as Nikkei reports, which could lead to equity dilution and a drop in investor confidence. That could make them even more vulnerable if the AI bubble does pop, or the industry fails to generate enough demand to justify the data center construction frenzy. The pressure is on: four of the five companies Nikkei analyzed are set to report second quarter earnings in the coming days and weeks. We'll be watching.

Chris Wood warns of massive capital destruction in US as China challenges AI boom

Jefferies strategist Chris Wood warns that the AI infrastructure boom could trigger massive capital destruction as cheaper Chinese open-source models challenge US dominance. With hyperscalers set to spend nearly $1.6 trillion over two years, investors are increasingly questioning whether returns will justify the unprecedented capex and rising debt. Jefferies' Global Head of Equity Strategy Chris Wood has warned that hundreds of billions of dollars being poured into AI infrastructure could culminate in "massive capital destruction" as cheaper Chinese open-source models erode the economics underpinning America's investment frenzy. Wood's longer-term base case is that market share will shift towards Chinese large language models, while investors increasingly question whether US technology companies can generate adequate returns from their unprecedented capital expenditure. Microsoft, Alphabet, Amazon and Meta are expected to spend a combined $695 billion on capital expenditure in 2026, rising to $870 billion in 2027, according to figures cited in Wood's latest GREED & fear report. Together, that amounts to nearly $1.57 trillion over two years. Alphabet raised its 2026 capital expenditure guidance by another $15 billion to between $195 billion and $205 billion. Investors will now focus on the guidance from Microsoft, Amazon and Meta as they report earnings. The scale of spending has transformed businesses once known for their asset-light models. After raising their guidance in April, the four hyperscalers' estimated capital expenditure reached an "astonishingly high" 92% of their forecast operating cash flow for 2026, according to Wood. The market initially welcomed that spending, partly because surging revenue at AI companies appeared to validate demand. Anthropic's annualised revenue run rate jumped from $9 billion in December to $47 billion in May, reinforcing optimism about corporate adoption and the monetisation of agentic AI. But Wood said investors are now beginning to ask the question that had largely been deferred: where will the returns on this capital come from? Also Read | Chris Wood's big warning: The specific risk that will finally trigger the end of AI trade China challenges US dominanceThe threat from China is no longer confined to cheaper models with Wood saying that there is a growing realisation that China has become a technological peer to the US in artificial intelligence. The top Chinese AI models processed 36.39 trillion tokens on OpenRouter during the week ended July 19, compared with 7.39 trillion tokens for the leading US models. Chinese models, therefore, handled nearly five times as many tokens on the global aggregation platform. Competition intensified with the July 17 launch of Moonshot AI's open-source Kimi K3. The model was estimated to offer about 95% of the performance of Anthropic's Claude Fable 5, according to the data cited in the report. The development builds on the "DeepSeek moment" of January 2025, which first brought the cost advantage of Chinese open-source models and the associated commoditisation threat to global investors' attention. Wood's concern is that the continuing decline in token prices could prevent large language models from becoming sustainably profitable. The Silicon Data LLM Token Expenditure Index, which tracks the average price paid for one million AI tokens, has fallen 25% since its late-May peak to $1.55. Falling prices may stimulate long-term demand for computing power, but they also threaten the profitability assumptions behind the current investment cycle. Also Read | Christopher Wood warns of AI fatigue. Why Jefferies is turning to India and China AI boom acquires a credit dimensionThe risk is no longer confined to equity valuations. AI infrastructure spending has increasingly been financed with debt rather than the hyperscalers' cash, giving the boom a growing credit-market dimension. The leading hyperscalers have raised $194 billion through investment-grade debt in 2026, making them the largest single source of issuance and putting them well ahead of the US energy sector's $55 billion. Credit markets are beginning to show signs of concern. The spreads on 10-year bonds issued by Amazon, Alphabet and Meta have widened to 78, 70 and 104 basis points over US Treasuries, respectively, from 61, 57 and 87 basis points on July 3. Oracle, a more leveraged participant in the AI infrastructure race, had its $120 billion debt pile downgraded to BBB- on July 9, leaving it one notch above junk status. Its 10-year bond spread has widened from 176 basis points to 219 basis points since the downgrade. The larger vulnerability lies in the revenue backlogs being used to justify infrastructure investment. Microsoft, Alphabet, Amazon and Oracle had about $2.1 trillion of remaining performance obligations at the end of the first quarter of 2026. These represent contractual commitments for future revenue and have surged 184% from $740 billion a year earlier. About half of that backlog is owed by OpenAI and Anthropic, according to figures cited by Wood. Microsoft's backlog has about 49% exposure to the two AI companies, while the corresponding exposure is 54% for Oracle, 43% for Google and 51% for Amazon. Neither OpenAI nor Anthropic is currently profitable, although Wood said Anthropic appears more comfortably positioned. The concentration means hyperscalers have effectively extended large, unsecured commitments to cash-burning customers while building data-centre capacity on the assumption that future computing demand will materialise. The risks are even greater for specialised cloud providers that have themselves borrowed to finance chips and infrastructure. CoreWeave has borrowed about $30 billion, while its five-year credit-default-swap spread has climbed from 452 basis points in early June to 701 basis points. Hidden liabilities and flattering earningsThe balance-sheet risks may also be understated. Wood cited an estimate that the five leading US hyperscalers had accumulated $662 billion of future data-centre lease commitments that had not yet commenced, up from $152 billion at the end of 2023. A separate study put their off-balance-sheet or "hidden" debt at $1.65 trillion in the latest quarter, exceeding the roughly $1.35 trillion of debt reported on their balance sheets. At the same time, the accounting impact of the investment wave has yet to catch up fully with the expenditure. Microsoft, Amazon, Alphabet and Meta collectively spent $130 billion on capital expenditure in the first quarter, while recording depreciation and amortisation expenses of $41.6 billion. Depreciation was nevertheless 33% higher from a year earlier. Wood also flagged what he described as financial engineering in the hyperscalers' recent profit growth. Their annualised earnings increased by $106.6 billion from a year earlier to $447 billion in the four quarters through March. Other non-operating income increased by $71.5 billion to $83 billion, accounting for about two-thirds of the earnings increase. The recent earnings strength could therefore be obscuring the scale of the risks accumulating beneath the AI trade. Wood stressed that AI is not a passing story and that falling computing costs could ultimately drive much greater usage. His warning is about timing: markets may have overestimated the technology's near-term returns while underestimating the capital and credit risks required to reach its long-term potential. "The time for an extended AI hangover after the initial surge of enthusiasm is approaching, if it has not already arrived," Wood said. (Disclaimer: Recommendations, suggestions, views and opinions given by the experts are their own. These do not represent the views of Economic Times)

AI investment boom faces growing credit risks as hyperscalers ramp up debt: Jefferies

Investors are now scrutinising massive artificial intelligence infrastructure spending by major tech companies. Hyperscalers are increasingly relying on debt to finance these significant AI investments. Upcoming earnings reports will keep AI capital expenditure under close scrutiny by investors. Contractual commitments for AI infrastructure have significantly increased over the past year. It remains uncertain which companies will successfully monetize their AI investments long-term. The rapid expansion of artificial intelligence (AI) infrastructure is entering a new phase as investors begin to focus on rising credit risks linked to massive spending by major US technology companies, according to a Jefferies report, which said hyperscalers are increasingly relying on debt to finance AI investments while questions over future returns are beginning to intensify. The report said the upcoming earnings season for Microsoft, Amazon and Meta will keep AI capital expenditure under close scrutiny after Alphabet recently increased its 2026 capital expenditure guidance by another USD 15 billion to USD 195-205 billion. Combined capital expenditure by the four major hyperscalers is estimated at about USD 695 billion in 2026 and USD 870 billion in 2027, highlighting the scale of the AI investment cycle. "If this is the backdrop it seems to GREED & fear that the questioning of the returns on AI capex... has now begun and could well intensify in the forthcoming earnings season," the report said. According to Jefferies, one of the biggest changes in the AI investment cycle is that hyperscalers are now funding a growing share of their spending through debt rather than cash. The report said the companies have issued USD 194 billion of investment-grade debt so far this year, making them the largest issuer in the US investment-grade debt market, while bond yield spreads for companies such as Amazon, Alphabet and Meta have widened. The report also highlighted growing contractual commitments linked to AI infrastructure, noting that remaining performance obligations (RPOs) across Microsoft, Amazon, Alphabet and Oracle reached about USD 2.1 trillion at the end of the first quarter of 2026, up 184 per cent from a year earlier. It said these commitments have largely been viewed as evidence of demand supporting AI-related data centre expansion, but added that investors may increasingly examine the associated credit risks as the industry evolves. Looking ahead, Jefferies said it remains uncertain which hyperscalers will successfully monetise their AI investments over the long term. At the same time, it noted that AI remains a long-term growth theme, saying demand for computing power is expected to continue rising even if investors become more selective about the pace and financing of future AI spending.