CoreWeave secured a delayed-draw term loan facility of up to $8.5 billion to scale its AI cloud infrastructure. The initial draw is about $7.5 billion with an option to increase to $8.5 billion as data center assets stabilize.
The seven-year loan, maturing in March 2032, was arranged by Morgan Stanley and MUFG and anchored by Blackstone Credit’s Insurance. CoreWeave says this milestone is a part of the $28 billion raised in 12 months that reflects strong market confidence in AI demand. Cloud specialist CoreWeave will use the funds to deliver on major AI contracts and accelerate its infrastructure build-out. “We’re proud to partner with leading financial institutions on this landmark transaction… This reflects confidence in AI adoption and market validation of our model,” said Brannin McBee, CoreWeave co-founder. The loan includes a SOFR-based floating tranche at SOFR+2.25% and a fixed-rate tranche at about 5.9%. Specific covenants were not disclosed.
READ: Blackstone gets rid of Fidere Residential portfolio (March 31, 2026)
CoreWeave completed an IPO in March 2025 and has rapidly expanded, including a recent UK data center investment and a reported 18% share of the dedicated AI GPU market. The financing comes amid a capital spending boom on AI infrastructure. Bank of America and Reuters note that U.S. data center investment hit record highs as tech giants pour billions into AI. CoreWeave is competing with both hyperscalers’ cloud offerings and smaller GPU-focused players. For example, Lambda Labs raised $480 million in early 2025 and secured a $500 million GPU-backed loan. Crusoe Energy recently closed a $350 million Series C and secured $200 million in asset-backed financing.
READ: SoftBank secures a $40 billion loan to boost investments in OpenAI (March 27, 2026)
High leverage is a risk if AI demand slows or supply chains disrupt GPU deliveries. CoreWeave will need to deploy equipment rapidly to service contracts and refinance debt as it expands. Next steps include drawing on the facility over the coming quarters to fund data center builds and chip purchases. Its progress will be watched against rivals’ moves and the broader AI cycle.