CoreWeave is changing how it pays for expansion as demand for AI computing grows. The company is asking some customers for money up front, borrowing against specific contracts and directing more capacity toward shorter-term deals that can carry higher prices.
Customer payments help cover expansion
CoreWeave rents computing power used for AI. Its third-quarter results showed $3.66 billion in revenue, up 168% from the year prior, alongside a $709 million net loss and $846 million in quarterly interest expense.
The company expects about $9 billion in customer prepayments and deposits over the next three months. More than $3 billion of that is tied to new customer contracts for capacity CoreWeave has already built, so those agreements do not require additional hardware spending.
Prepayments can give CoreWeave cash before it needs to borrow for more infrastructure. But they also commit the company to delivering computing capacity at agreed prices. If demand keeps rising, those prices could look less attractive later.
CoreWeave says its backlog stands at $132.5 billion. To build the infrastructure needed to serve it, the company plans to use loans backed by particular customer contracts and draw funds in installments as needed. CEO Michael Intrator said, “We will not use the unsecured debt market to fund our business at current yields, nor do we need to.”
Shorter contracts can bring higher prices
CoreWeave is also allocating more capacity to short-term opportunities. These deals can let the company set prices more frequently, and the company says some short-term capacity is selling for more than $40 million per megawatt annually.
By comparison, pricing for new infrastructure on four and five-year contracts is moving to $20 million to $25 million per megawatt. Longer agreements may bring lower prices, but they offer more predictable revenue over time.
CoreWeave raised prices across its offerings by about 25% in July and by another 10% later in the quarter. Intrator said prices are still climbing this quarter in a way that “more than offset[s] cost increases.”
That pricing gives the company a reason to pursue shorter deals, but it also makes its financing approach more complicated. Lenders tend to place greater weight on longer-term contracts because consistent cash flow can make repayment easier to predict.
Contract-backed borrowing shifts the tradeoffs
In August, CoreWeave closed a $2.6 billion loan designed specifically for shorter-term customer contracts. Called DDTL 5.5, the facility lets the company borrow in installments rather than taking all the money at once.
Borrowing against contracts ties financing to specific customer arrangements. That can help CoreWeave avoid relying on unsecured debt, which Intrator said the company would not use at current yields. Still, shorter contracts may be less reassuring to lenders than long agreements, even when short-term pricing is higher.
Other companies in the sector are also using customer prepayments to fund computing capacity. Nebius and IREN have relied on hefty Microsoft prepayments, pointing to a broader move toward asking Big Tech customers to help pay for infrastructure.
Demand and investor confidence remain central
CoreWeave’s strategy depends on customers continuing to need large amounts of AI computing and on the company being able to charge prices that support its costs. Shorter contracts could allow it to respond as prices change, but they bring more uncertainty when agreements expire and need to be renewed.
Investors have also shown different views of the company’s financing risks. Wall Street reacted positively to the earnings, while yields on some CoreWeave bonds hit above 15% last month. Those yields highlight concerns about the debt needed to build data centers before the investment produces a return.
Customer cash and contract-backed loans offer CoreWeave another way to expand while managing its borrowing. Whether that approach pays off will depend on sustained demand, attractive pricing and the company’s ability to keep its capacity under contract.