Global banks face limited CoreWeave exposure, J.P. Morgan says

Traders Union
All News
Financial Institutions
Global banks' CoreWeave exposure

Highlights

J.P. Morgan estimates large U.S. investment banks and MUFG have CoreWeave exposure of $1 billion to $3 billion, with most others below $500 million.
CoreWeave holds about $24 billion in committed delayed draw term loans—$15 billion drawn and $9 billion undrawn—within $50 billion total debt.
Yield on CoreWeave’s notes due 2031 hit nearly 16% last week, as J.P. Morgan projects negative free cash flow until 2030 amid heightened lender risk.

Bank exposure estimates and loan structure

As reported by Reuters, citing J.P. Morgan, most global banks have limited exposure to CoreWeave, even as the Nvidia-backed cloud infrastructure company remains heavily financed through loans and bond markets. The brokerage estimated that large U.S. investment banks, along with Japan's MUFG, could have exposure of about $1 billion to $3 billion, while most European and Asian banks have less than $500 million.

J.P. Morgan said much of that exposure comes through delayed draw term loans, which are largely held through special-purpose vehicles and release capital as projects meet milestones. Analysts led by Kian Abouhossein said bank exposure to CoreWeave appears well managed and very well secured, while noting that the estimates rely on simplified assumptions because banks disclose limited detail about their links to the AI investment cycle.

The brokerage also said CoreWeave has about $24 billion of committed delayed draw term loans, including $15 billion already drawn and $9 billion still undrawn. J.P. Morgan maintains an overweight rating on the company with a target price of $125 for the end of 2027.

AI financing risks and broader market impact

Investor unease is spreading across AI infrastructure markets as technology companies raise money aggressively from private credit investors, banks and bond markets to expand capacity quickly. That funding wave is increasing scrutiny over whether returns will justify the scale of spending.

CoreWeave, which provides high-performance computing infrastructure for AI, has secured tens of billions of dollars in contracts with major technology companies and carries about $50 billion in debt. Last week, the yield on its notes due 2031 reached nearly 16%, while the cost of insuring against its debt rose to around 1,000 basis points.

J.P. Morgan said the main risk for lenders is a mismatch between large upfront spending on AI infrastructure and revenue that arrives over time. The bank expects CoreWeave's free cash flow to remain negative until 2030, underscoring the longer payback period facing the sector.

In our earlier coverage of the U.S. AI data center buildout, we noted that the boom is running into higher financing costs, construction delays, and growing power constraints that are testing investor confidence. The report highlighted that only a minority of planned projects appear credible and pointed to signs of stress across the sector, including CoreWeave’s sell-off and cost overruns at other developers—raising questions about whether massive upfront spending can translate into sustainable returns.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.