CoreWeave reported third-quarter revenue of $3.661 billion on November 10, up from $1.365 billion a year earlier, while a larger interest bill weighed on its bottom line. The results put a familiar infrastructure question in focus: how quickly expanding AI capacity can be financed as the business grows.
The company’s earnings release showed operating income of $207 million and a net loss of $709 million for the quarter ended September 30. Interest expense reached $846 million. Those figures distinguish the operating performance of the cloud business from the financial cost of building and funding its capacity.
Power and contracts underpin expansion
CoreWeave said it added more than 200 megawatts of active power during the quarter, bringing the total above 1.7 gigawatts. Contracted power exceeded 4.6 gigawatts. The company reported revenue backlog of $132.5 billion, a measure that includes commitments whose delivery depends on contractual conditions.
For 2026, management guided to revenue of $13 billion to $13.3 billion and capital expenditures of $37 billion to $39 billion. These are forecasts, not completed results. The scale of planned spending makes the source and timing of financing a central part of the expansion story, alongside customer demand.
Financing tied to equipment and customers
In CoreWeave’s third-quarter earnings-call transcript, management described a strategy centered on delayed-draw loans linked to particular customer contracts. These facilities are drawn as equipment is purchased and delivered. The company said it did not intend to issue new parent-level unsecured debt through 2027 at current market yields.
Management said more than three-quarters of committed capital spending through 2027 was associated with investment-grade or enterprise customers. It described loan structures covering between 70% and 100% of the relevant capital expenditures, with contract cash flows supporting debt service. These figures describe management’s financing approach rather than eliminating execution or credit risk.
The transcript also outlined expected customer deposits and prepayments of approximately $9 billion over the next three months. CoreWeave said prepayments featured in 70% of deals signed this year. The important distinction is that cash received ahead of service delivery helps funding, but is not the same as revenue already earned.
What the figures do and do not show
The results illustrate why AI infrastructure cannot be assessed through revenue growth alone. A provider may add sales and operating income while still recording a net loss after financing costs. Equally, contracted capacity and backlog describe future business opportunities, not proof that every facility is already operating or every dollar has been collected.
For customers evaluating infrastructure suppliers, the practical questions are about delivery as much as scale. Contract milestones, available power, equipment installation and funding arrangements need to align. A large order book can support expansion plans, but the work of turning commitments into usable computing capacity remains separate.
The next checkpoints are whether capacity arrives on schedule and whether the funding arrangements described by management close on the expected terms. CoreWeave’s report offers evidence of rapid growth and substantial committed demand. It also shows why financing costs and the timing of cash flows remain essential to understanding that growth.
