Fed’s first close look at AI buildout finds $3 trillion of commitments off Big Tech’s balance sheets
The Fed's November Financial Stability Report highlights over $3 trillion in AI-related commitments from major tech firms, including leases and contracts. These obligations are off-balance-sheet but could pose risks if AI demand falls short. The report notes that the seven largest AI firms' on-balance-sheet debt is expected to exceed $800 billion by year-end.
How this was made

The 30-second read
Why it matters
The article frames a potential tail-risk channel: if AI demand disappoints, losses may be absorbed by developers, private debt vehicles, and parts of the financial system, with opacity potentially amplifying selloffs.
Market read
Traders may reprice AI infrastructure credit risk and off-balance-sheet leverage, especially into Nvidia’s earnings given the article’s direct linkage to Nvidia guarantee structures.
What to watch
The Fed notes off-balance-sheet figures are undiscounted and not directly comparable to debt, and the banking system is described as sound and resilient; the real impact may depend on how much of these commitments translate into realized losses.
Background
The Fed’s Financial Stability Report adds a first-time section quantifying leverage in the AI data-center buildout, focusing on signed leases and take-or-pay/residual value guarantees that sit off balance sheet.
Ticker impact
Fed report highlights Nvidia’s residual value guarantees and lease guarantees tied to AI data centers, implying off-balance-sheet risk if demand falters.
Likely modest downside bias as investors reprice off-balance-sheet leverage and counterparty default risk around the AI buildout.
The article does not announce a new Nvidia-specific financial change, but it surfaces Fed-quantified leverage and cites Nvidia’s guarantee mechanics and caps, which can affect risk perception into its earnings call.
Fed’s new section counts more than $2T in take-or-pay contracts and residual value guarantees at Amazon as part of the AI data-center buildout.
Likely neutral-to-negative sentiment as markets consider who absorbs losses if AI leasing demand weakens.
The Fed framing is broad and not a new Amazon contract, but it explicitly includes Amazon among the largest AI firms with large commitments.
Fed report includes Alphabet among seven largest AI firms with on-balance-sheet debt expected to exceed $800B, plus large off-balance-sheet commitments.
Slight negative bias as investors price opacity and potential selloff amplification from hidden leverage.
The article provides new Fed-reported totals and explicitly names Alphabet, but it does not provide a new Alphabet-specific filing or guidance.
Fed’s leverage-in-data-center-buildout section counts large signed-but-unstarted leases and take-or-pay/residual value guarantees at Broadcom.
Neutral-to-negative as investors reassess AI buildout financing risk across the supply chain.
Broadcom is named, but the article does not detail Broadcom-specific guarantee terms like it does for Nvidia.
Fed’s new section explicitly includes Meta in the $3T+ commitments off Big Tech balance sheets tied to AI data-center buildout.
Likely neutral with potential downside if markets extrapolate default risk into AI capex financing.
Meta is included in the Fed totals, but there is no Meta-specific new contract detail in the article.
Fed report counts more than $2T in take-or-pay contracts and residual value guarantees at Microsoft as part of AI data-center financing.
Slight negative bias as investors focus on opacity and selloff amplification from off-balance-sheet obligations.
The article provides new Fed framing and totals, but no Microsoft-specific incremental disclosure beyond being named.
Fed’s leverage section includes Oracle among firms with large off-balance-sheet AI data-center commitments.
Neutral-to-negative as risk appetite toward AI infrastructure exposure cools.
Oracle is named in the Fed totals, but the article lacks Oracle-specific guarantee mechanics or new filings.
Market effects
Highlights credit and liquidity risk in AI data-center financing, potentially pressuring hyperscalers, GPU supply chain, and private-credit sentiment.
Primarily US financial-system risk framing, with spillover to US bank and private-credit exposures.
AI infrastructure buildout financing is global, but the catalyst here is US Fed analysis and US bank exposure.
Counterpoint
Barr explicitly says he is less concerned about direct losses given limited debt backing the current boom, implying the market may overreact to off-balance-sheet opacity.
Key entities
- regulatorFederal Reserve Financial Stability Report
Introduces a new section, “Leverage in the Data Center Buildout,” quantifying AI-related commitments off Big Tech balance sheets.
- officialMichael Barr
Fed Governor who cited the report in a speech in Milan and warned opacity can exacerbate selloffs.
- companyNvidia
Named among firms with large AI-related guarantees; article details Nvidia’s residual value guarantees and caps tied to an AI campus and cloud partners.
- companyOpenAI
Referenced as a tenant in the Ohio AI campus arrangement and as the party that would reimburse Nvidia if Nvidia pays under guarantees.




