CoreWeave Shifts to Prepayments and Shorter Contracts to Fund AI Infrastructure

Photo credits
CoreWeave is altering how it finances its artificial intelligence infrastructure, relying increasingly on customer prepayments, contract-backed loans, and shorter-term agreements rather than traditional unsecured debt. The neocloud provider rents out computing power for AI workloads, and its updated strategy aims to fund growth efficiently while extracting more cash from its existing infrastructure.
During its third-quarter earnings report on Tuesday, the company posted $3.66 billion in revenue, representing a 168% increase from the prior year. However, that growth was accompanied by a $709 million net loss and $846 million in quarterly interest expenses.
“We will not use the unsecured debt market to fund our business at current yields, nor do we need to,” CoreWeave CEO Michael Intrator said during the earnings call.
Contract-Backed Funding and Prepayments
Instead of issuing unsecured debt, CoreWeave plans to rely on loans tied to specific customer contracts, drawing down funds as needed to build out the infrastructure required for its $132.5 billion backlog. The company expects to receive approximately $9 billion in customer prepayments and deposits over the subsequent three months. This total includes over $3 billion associated with new customer contracts for computing capacity that has already been constructed, requiring no additional hardware expenditures.
Other market participants, including competitors Nebius and IREN, have similarly utilized substantial Microsoft prepayments to fund computing capacity, indicating a broader industry trend toward requiring major technology companies to finance their own buildouts.
While upfront payments reduce immediate borrowing requirements, they bind the company to service delivery at fixed prices. Should computing demand continue to surge, those fixed prices may become less financially advantageous over time.
Short-Term Deals and Pricing Pressures
To capture value from rising demand, CoreWeave is allocating more capacity to shorter-term deals that command higher rates. The company increased its service prices by approximately 25% in July, followed by another 10% increase later in the quarter, with pricing continuing upward.
“We are also allocating more capacity to short-term opportunities, where the pricing justifies it, selling capacity at more than $40 million per megawatt annually, while pricing for new infrastructure on four and five-year contracts is moving to $20 million to $25 million per megawatt,” Intrator said.
This operational focus prompted the company to close a $2.6 billion loan in August, designated as DDTL 5.5, which is specifically tailored for shorter-term customer contracts and allows borrowing in installments. Although short-term agreements enable higher and more flexible pricing, they present a financing trade-off. Lenders generally favor longer-term contracts because they offer predictable cash flows and consistency for loan repayment, even when priced lower.
While equity markets reacted favorably to the earnings announcement, bond investors expressed caution. Yields on certain company bonds climbed above 15% in the preceding month, reflecting investor anxiety regarding the debt load required to finance data center construction prior to realizing a return on those investments.
Summary
CoreWeave’s transition toward customer prepayments and short-term agreements illustrates how AI infrastructure providers attempt to finance massive capital expenditures while avoiding high-yield unsecured debt. While upfront cash infusions and higher-priced short-term deals can boost near-term cash flow, they also increase exposure to contract renewal risks and future shifts in market demand.