Debt issued to finance data centers and other advanced computing infrastructure rose to more than $680 billion in 2026 through October, the Federal Reserve Board said in the Financial Stability Report it released on November 13. That is about five times the $130 billion issued in all of 2023. The borrowing’s share of nonfinancial investment-grade corporate debt issuance went from less than 6 percent in 2023 to about 31 percent in 2026.

The figures appear in a box titled “Leverage in the Data Center Buildout,” which was not in the Fed’s May report; the report generally reflects information available as of October 28. The issuance totals come from Bank of America Global Research and Bloomberg data and include public deals and trackable private securities, but exclude private placements under section 4(a)(2). The Fed estimates outstanding public debt related to advanced computing at more than $1.1 trillion as of October and says sizable private credit deals make the full amount harder to measure.

Leases, guarantees and other obligations have grown faster than the debt, according to Fed staff estimates drawn from the filings of Amazon, Alphabet, Broadcom, Meta, Microsoft, Nvidia and Oracle. Leases that have not yet commenced exceed $1 trillion, and take-or-pay contracts, residual value guarantees and similar obligations exceed $2 trillion. Both figures are undiscounted, and the report says they are not directly comparable with discounted on-balance-sheet debt. Microsoft, Meta, Alphabet and Amazon reported about $1.02 trillion in leases not yet started as of September 30.

Hyperscalers often sign long-term leases with third-party data center developers, who use the leases to secure construction financing, the report says. GPU makers often guarantee the residual value of GPU collateral to support purchases of their chips by advanced computing and large language model companies. Those financings are often supplemented with equity investments by the same hyperscalers and GPU designers, which the Fed says creates a “circular financing” dynamic.

The Fed wrote that through these links the largest technology firms may absorb losses in a future downturn, limiting the effects on the rest of the financial system. Their economic exposure, it added, “may be substantially larger than suggested by their on-balance-sheet liabilities.” The riskier subordinated debt and equity tranches of the special purpose entities used in many of these financings are held by high-yield bond funds, life insurers, business development companies and other private debt vehicles.

The largest US bank holding companies have close to $1 trillion of exposure to large technology firms, data center developers, advanced computing-related special purpose entities and private credit vehicles, and utilities. Those exposures are evaluated in the annual supervisory stress test, the report says.

In the New York Fed’s fall survey of 23 market contacts, conducted in September and October, AI was the most cited risk to financial stability, with respondents pointing to equity valuations, AI-enabled cyberattacks and complex AI-related financing. In the spring survey of 20 contacts, AI ranked third, behind geopolitical risks and an oil shock. The Fed says the survey reflects respondents’ views, not its own.

The report’s cyber section says reports of agentic systems being involved in cyber events have increased in frequency since May, amid an overall rise in reported cyber events. It says advances in the ability of large language models and agentic systems to detect and exploit vulnerabilities have created new challenges in securing financial institutions, financial infrastructure and third-party service providers.

Sources: Federal Reserve Board

–
By the Control Plane Editorial Team