Amid the frenzy of AI infrastructure construction, an unprecedented debt expansion is quietly taking shape—with its most dangerous portion never appearing on any balance sheet.
Goldman Sachs's latest report predicts that capital expenditures for hyperscale cloud companies will reach $1.1 trillion to $1.4 trillion by 2027, far exceeding market consensus. However, according to an in-depth study by Morgan Stanley, this already staggering figure is just the tip of the iceberg.
Nearly $1 trillion in purchase commitments, over $800 billion in unactivated lease contracts, and tens of billions of dollars in supplier financing arrangements collectively constitute an off-balance-sheet exposure of approximately $1.8 trillion—these liabilities exist outside the balance sheet but genuinely lock in future cash outflows.
The market has not yet fully priced in these risks.
Morgan Stanley warns that the leverage ratio of hyperscale cloud companies has surged from 0.9x to 1.8x in just two quarters, with capital expenditure growth continuing to outpace revenue and free cash flow growth, while the real impact of depreciation pressure has yet to arrive.
Meanwhile, private credit firms represented by Apollo and Blackstone are transferring leverage to the supply chain level through SPVs (Special Purpose Vehicles), creating a highly circular and difficult-to-penetrate financing structure. If AI commercialization falls short of expectations or enterprise customers shift en masse to cheaper alternatives, the fragility of the entire financing chain will be exposed.
Debt Issuance Frenzy: AI Has Become the Biggest Variable in Public Markets
According to Morgan Stanley's latest "AI Debt Financing Tracker Report," as of the end of May 2026, the scale of global AI-related bond issuance has reached $236 billion, a surge of 357% compared to the same period in 2025.
Morgan Stanley expects the total issuance of AI debt to exceed $570 billion for the full year, with the pace accelerating further in the second half of the year as capital expenditure financing needs are concentrated.
In April alone, AI-related bond issuance exceeded $74 billion, hitting a new high for the year, with project financing structures (for data center construction) accounting for 85% of high-yield bond supply and 40% of investment-grade bond supply. Meanwhile, the five hyperscale cloud companies—Amazon, Meta, Google, Microsoft, and Oracle—now account for 4% of the entire investment-grade bond index.
In terms of leverage, the gross leverage ratio for hyperscale cloud companies has risen from 0.9x in the third quarter of 2025 to the current 1.8x, increasing by approximately 0.3x per quarter, already surpassing the leverage level of the entire energy industry.
Morgan Stanley points out that, affected by supply pressure, related credit spreads have drifted from the AA range to the A range and may widen further. Meta's credit spread is currently wider than the CDX IG benchmark.
Regarding free cash flow, Morgan Stanley predicts that Amazon and Meta's free cash flow in 2026 will approach zero or even turn negative, at which point incremental financing will almost entirely rely on new debt.
$1.8 Trillion Off-Balance-Sheet Exposure: Invisible Liabilities, Locked-in Cash Outflows
Todd Castagno from Morgan Stanley's Global Valuation, Accounting & Tax team notes in the report that focusing solely on capital expenditure numbers would severely underestimate the true financial commitments of the AI construction cycle. Beyond the disclosed capital expenditures, there are three key types of off-balance-sheet exposure:
Purchase commitments of approximately $982 billion. The total value of long-term purchase contracts by hyperscale cloud companies and Nvidia is close to $1 trillion. Under accounting standards, these obligations are not recorded as liabilities until goods are delivered, unless the company expects a contract loss. Therefore, nearly $1 trillion in future cash outflows currently do not appear as liabilities on any balance sheet.
Notably, Nvidia's own inventory and purchase obligations have risen to about 32% of consensus revenue forecasts for fiscal year 2027, far above the historical range of 15% to 20%, indicating that supply chain commitment risks are extending to chip suppliers.
Unactivated lease commitments of approximately $822 billion. Over $800 billion in lease contracts have been signed but not yet activated and are not included in current lease liabilities. Additionally, arrangements such as variable lease payments, renewal options, and residual value guarantees also exist outside the balance sheet.
Morgan Stanley estimates that if finance leases were included in the calculation, Microsoft's capital expenditure-to-sales ratio would jump from 33%/50% (fiscal 2026/2027) to 44%/64%, while Oracle's could rise from 76%/115% to 101%/189%.
Unpaid capital expenditures in accounts payable of approximately $110 billion. The days payable outstanding (DPO) of hyperscale cloud companies have significantly lengthened—Oracle's increased by 370% year-on-year, Meta's by 73%, Microsoft's by 69%—meaning the entire supply chain is effectively financing AI construction, with suppliers bearing the liquidity pressure that should be carried by the buyers.
SPVs and Circular Financing: Leverage Moves to the Shadows
Another core dimension of off-balance-sheet risk is the circular financing structure built through SPVs.
A $35 billion "chip-collateralized" private credit deal completed this week by Apollo and Blackstone for Anthropic vividly illustrates the logic of this model:
Broadcom provides backing for this SPV; Anthropic uses the raised funds to purchase Google chips manufactured by Broadcom, with Google holding a 14% stake in Anthropic; Morgan Stanley, which arranged the deal, simultaneously provides loans to the investors participating in the transaction.
Morgan Stanley's AI ecosystem financing correlation map shows multiple circular relationships—customer, investor, supplier financing, and repurchase—among OpenAI, Oracle, Nvidia, Microsoft, CoreWeave, AMD, and Amazon. The same funds circulate repeatedly among a few key entities, with SPVs being the core tool enabling this circulation.
It is reported that Athene, the insurance subsidiary of Apollo, is particularly active in this structure—raising funds by selling annuities to retirees and then injecting the capital into SPVs to participate in AI infrastructure financing.
This model shifts leverage from the visible balance sheets of hyperscale cloud companies to suppliers and the private credit ecosystem, making the true systemic risk exposure difficult for external observers to identify and aggregate.
Depreciation Cliff and Monetization Gap: The Delayed Shock
Current financial data exhibits systematic optimism bias. A large amount of capital expenditure is currently booked as "construction in progress" (CIP) and has not yet begun to depreciate, artificially inflating reported profit margins and underestimating future expense pressure.
The CIP balances of Oracle, Meta, and Google have grown by approximately 200%, 90%, and 55% year-on-year, respectively.
Once these assets gradually transition to depreciation, the impact will be concentrated.
Morgan Stanley predicts that the cumulative depreciation for Microsoft, Oracle, Meta, and Google over the next three years will exceed $520 billion. Taking Oracle as an example, depreciation as a percentage of revenue could rise from the current 7% to 28% in fiscal year 2028; Meta's could rise from 9% to 19%.
In this context, the only path to maintaining profit margins is for revenue to grow significantly in sync—yet the upward revisions to revenue forecasts currently lag far behind the upward revisions to capital expenditure forecasts.
Data shows that consensus capital expenditure forecasts for Google for 2026 have been raised by 139% compared to a year ago, while Meta's and Amazon's were raised by 85% and 81%, respectively. Oracle saw the largest increase, at 175%.
Meanwhile, revisions to revenue forecasts are noticeably lagging, revealing a clear structural mismatch where capital expenditure is outpacing commercialization.
Furthermore, over $2 trillion in remaining performance obligations (RPO) is highly concentrated in a few large, long-term contracts, presenting significant counterparty concentration risk—if any major participant in this circular system encounters problems, it could trigger a chain reaction.
Timing Mismatch Rather Than an Immediate Solvency Crisis
Morgan Stanley concludes that these risks do not currently constitute an imminent solvency crisis but rather a series of overlapping timing mismatches and information disclosure gaps: depreciation pressure is deferred, capital expenditure outpaces monetization, leverage shifts to suppliers and the private credit layer, and the comparability of capital intensity between different companies is significantly undermined by accounting classification differences.
Hyperscale cloud companies are clearly aware of the limited window of current market sentiment and are seizing the opportunity to maximize financing scale.
Goldman Sachs analyst Ryan Hammond points out that if AI infrastructure investment scales to 2% to 3% of GDP, analogous to historical construction cycles of railways and the automotive industry, capital expenditures could reach $1.1 trillion by 2027. In an extreme scenario, considering the cash flow of hyperscale cloud companies and the capacity of the investment-grade credit market, the upper limit might reach $1.4 trillion.
However, all of this hinges on the ability of large language models (LLMs) to continue increasing token pricing and maintain sufficient enterprise customer stickiness. A growing number of enterprises are turning their attention to AI products with comparable performance but significantly lower prices.
Should a structural shift occur on the demand side, the meticulously constructed financing system of today will face a fundamental stress test.
