CoreWeave is rethinking how it pays for the AI boom, according to the company’s third-quarter results reported Tuesday. CoreWeave, the neo-cloud company that lends out the computing power behind AI, said it is increasingly asking customers to pay upfront or borrow under certain contracts, and even pursuing short-term deals that can lead to higher prices.
The new strategy could help CoreWeave more efficiently fund growth while squeezing more cash out of its existing infrastructure. But these bets only hold if today’s demand and pricing power holds.
CoreWeave reported revenue of $3.66 billion, up 168% year over year, but also posted a net loss of $709 million and quarterly interest expense of $846 million.
“We have no intention or need to use the unsecured debt market to fund our business at current yields,” Coreweave CEO Michael Intrater said on Tuesday’s earnings call.
Instead, Coreweave’s plan is to rely on financing backed by certain customer contracts to finance much of the infrastructure it still needs to build to meet its massive $132.5 billion backlog.
CoreWeave also said it expects to receive approximately $9 billion in customer prepayments and deposits over the next three months, including more than $3 billion in new customer commitments for capacity it has already built, which will not require additional spending on hardware.
Rivals Nebius and IREN also rely on large upfront payments from Microsoft to pay for their computing power, suggesting the industry is moving toward asking big tech companies to pay for their expansion.
While the upfront payment may ease CoreWeave’s immediate borrowing needs, it also commits the company to offering its services at prices that could become less attractive in the future if computing demand continues to soar.
CoreWeave is also looking to take advantage of that increased demand by entering into shorter-term contracts that allow it to command much higher prices.
The company raised prices on all its products by about 25% in July and another 10% later in the quarter. Intrater said prices continued to rise this quarter and “more than offset the cost increases.”
“We are allocating more capacity to short-term opportunities where pricing warrants, selling capacity for more than $40 million per megawatt per year, while new infrastructure prices on four- and five-year contracts are moving to $20 million to $25 million per megawatt,” the CEO said.
CoreWeave’s funding strategy currently revolves around these opportunities. In August, the company closed a $2.6 billion loan, its first loan specifically designed for short-term customer contracts. The tool, known as DDTL 5.5, allows CoreWeave to borrow money in installments when they need it.
Short-term trades allow CoreWeave to charge higher prices and change more frequently, but this is a trade-off. Lenders value long-term contracts even at lower prices because they value predictable cash flow and consistency in loan repayments above all else.
While Wall Street reacted positively to Coreweave’s earnings, bond investors were less optimistic. Yields on some of the company’s bonds rose above 15% last month, highlighting investors’ concerns about the debt needed to fund data center construction before they expect a return on their investments.
In other words, CoreWeave is finding new ways to grow its business without relying on unsecured debt. The strategy the company has chosen has the potential to increase profits, but it also comes with greater risks regarding contract renewals and pricing. And this shows how the AI infrastructure boom relies on sustained, unproven demand.
TechCrunch reached out to CoreWeave for more information about the loan facility.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
