CoreWeave Lands $8.5 Billion GPU-Backed Financing in Latest AI Deal

CoreWeave has closed an $8.5 billion financing facility tied to its AI infrastructure, marking what the company touted as the first investment-grade rated debt backed by GPU clusters and customer contracts.

CoreWeave announced the facilitiy on Tuesday, saying the deal is structured as a delayed draw term loan, which carries an initial borrowing capacity of about $7.5 billion, expandable to $8.5 billion as assets reach stabilization.

The debt was rated A3 by Moody’s and A (low) by DBRS Morningstar, a stage that underscores how capital markets are beginning to treat AI infrastructure more like traditional contracted assets such as power plants.

The loan is split between a floating-rate tranche priced at SOFR plus 225 basis points and a fixed-rate portion at roughly 5.9%, with maturity in March 2032. That implies an effective borrowing cost in the mid-6% range under current rate conditions—relatively lower than some of the high-yield debt deals that have funded much of the recent AI data center buildout.

The transaction was arranged by MUFG and Morgan Stanley, with Goldman Sachs and JPMorgan Chase as additional lead arrangers. Blackstone Credit & Insurance anchored the deal, which the company said was oversubscribed.

The financing is secured by assets held under a CoreWeave subsidiary and linked to long-term customer contracts, a structure that has become increasingly common as lenders seek predictable cash flows to underwrite large-scale compute infrastructure.

CoreWeave said it has now raised about $28 billion in equity and debt commitments over the past year, placing it among the most aggressive spenders in the race to build AI capacity.

The deal highlights a broader shift in how AI infrastructure is financed. Until recently, most data center developers and bitcoin miners relied on high-cost debt—often with coupons above 7%—or equity-heavy funding to expand capacity.

For CoreWeave, the lower cost of capital could prove decisive as competition intensifies among developers racing to secure power, land and chips. The company has positioned itself as a key intermediary between hyperscale demand and infrastructure delivery, building out GPU-heavy campuses designed for high-density workloads.

As AI demand continues to scale, the ability to translate long-term compute contracts into cheaper financing is emerging as a critical advantage—potentially reshaping not just who builds data centers, but who can afford to operate them at scale.