👩🍳 How we use AI at Tech in Asia, thoughtfully and responsibly.
🧔♂️ A friendly human may check it before it goes live. More news here
🧔♂️ A friendly human may check it before it goes live. More news here
OpenAI cloud supplier CoreWeave secures $3.1b loan
CoreWeave, a Livingston, New Jersey-based cloud provider for AI computing, has secured a US$3.1 billion loan backed by customer contracts for graphics processing units.
Pricing tightened to 4.5 percentage points above the benchmark after investor orders topped US$15 billion.
The loan is offered at 99 cents on the dollar.
The 5.5-year delayed-draw term loan will fund GPU purchases and installation for OpenAI and Cohere.
It is CoreWeave’s first GPU-backed financing to be syndicated in the US leveraged loan market.
It follows the company’s US$8.5 billion chip-backed financing in March.
🔗 Source: Bloomberg
🧠 Food for thought
Implications, context, and why it matters.
A new loan highlights a shift to riskier bets
- The loan is set for a credit rating in the “BB” range, which is below investment grade. It rests on customer contracts with OpenAI and Cohere, which are unrated and seen as more speculative counterparties 1.
- CoreWeave’s March financing carried investment-grade ratings because the customer contracts behind it came from Meta, the parent company of Facebook and Instagram 2.
- The pricing tracks that added risk. The March delayed-draw term loan, or a loan borrowers can access over time, included a floating-rate tranche at Secured Overnight Financing Rate (SOFR) + 2.25%. Talks on the new loan started at SOFR + 5.0%, then moved to about SOFR + 4.5% after demand picked up 1.
The AI boom is being built on a mountain of high-risk debt
- This deal fits a wider pattern. Debt issuance for data centers reached US$183 billion in 2025, up from US$92 billion a year earlier, according to S&P Global, a financial data and ratings firm 3.
- CoreWeave carries heavy leverage. It had about US$15.2 billion in interest-bearing debt against US$3.34 billion in shareholders’ equity, or about 4.5x. Total liabilities, including leases, were roughly US$46 billion 4.
- These loans depend on collateral in the form of GPUs, the chips used to train and run AI systems. Their useful economic lives are disputed, may be shorter than some depreciation schedules, and rental rates have dropped sharply from 2023 levels 5.
- A default could spread. It may set off cross-defaults across CoreWeave’s credit facilities, hit several lenders, and ripple through securitized and off-balance-sheet financing tied to the sector 5.
Recent CoreWeave developments
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.




