The Risk Behind the AI Buildout

• Banks are trying to transfer billions of dollars of AI-infrastructure credit risk to insurers while freeing themselves to make even more loans.
• Barclays is reportedly seeking insurance protection on more than $10 billion in AI-infrastructure debt, while HSBC and Deutsche Bank are pursuing their own risk-transfer strategies.
• Some credit insurers are already approaching their limits on data-center exposure, raising a larger question about who ultimately carries the financial risk of the AI boom.

NEW YORK - The enormous AI data-center buildout has created an unusual new role for insurance companies - Wall Street wants them to absorb part of the risk if billions of dollars in AI infrastructure debt goes bad.

HSBC is in preliminary discussions with insurers about transferring credit risk on project-finance loans, including financing tied to AI infrastructure and data centers, according to Insurance Journal’s October 30 report. HSBC has also been working on a significant risk transfer, or SRT, involving a portfolio of more than $1 billion in data-center-related loans.

HSBC says its exposure is relatively small. During the bank’s October 28 investor call, Chief Financial Officer Pam Kaur said data-center lending represents only a small portion of HSBC’s business in both the United States and Asia. The official HSBC earnings-call transcript confirms her remarks.

HSBC is not alone.

Barclays is reportedly seeking insurance protection on more than $10 billion of debt financing AI infrastructure. On some facilities, the bank reportedly wants insurers to assume more than half of the credit risk. The discussions remain exploratory, and no final transaction has been announced. Insurance Journal reported on Barclays’ discussions here.

Deutsche Bank, meanwhile, has been working on a significant risk transfer involving approximately €2 billion, or roughly $2.3 billion, in project-finance debt that includes data-center loans.

This is where the insurance story becomes much bigger than three banks.

Credit-risk insurance allows a lender to pay an insurer to cover an agreed portion of losses if borrowers default. By transferring qualifying risk, a bank can potentially reduce the amount of regulatory capital tied up against those loans. That can create room on the bank’s balance sheet to lend again.

In other words, insurance may not simply protect banks from the AI boom. It may help finance the next stage of it.

But insurance capacity is not unlimited.

WTW’s 2026 Credit and Political Risk Insurance Capacity Survey identified rapidly growing insurer exposure to data-center financing and increased use of insurance for bank de-risking structures. Reporting on the market says several providers have indicated they are nearing—or have already reached—their sector limits for data-center exposure.

The numbers explain why.

McKinsey estimates that data centers could require roughly $6.7 trillion in worldwide capital investment by 2030, with approximately 70% of the demand associated with AI workloads.

That creates an important distinction for investors and policyholders. The risk being transferred here is not primarily the familiar property risk of a data center catching fire or suffering equipment damage. It is credit risk—the possibility that borrowers cannot repay enormous infrastructure loans on the terms originally expected.

The unresolved question is then not whether AI will require more computing power. It clearly will.

The harder question is whether the economic returns from that infrastructure will ultimately justify the debt being accumulated to build it.

Banks originate the loans. Technology companies consume the computing capacity. Investors provide capital.

And increasingly, insurance companies are being asked to stand behind the debt.

If that trend accelerates, one of the most consequential risks created by the AI spending boom may quietly migrate from bank balance sheets onto insurer balance sheets—precisely as some insurers are signaling that their capacity is already being tested.

Enjoy reporting that goes beyond the headline? Subscribe to the USA Herald newsletter for investigative reporting, legal and insurance analysis, national-security coverage, and emerging stories that deserve a closer look.

👉 Follow USA Herald on X at @RealUSAHerald