CoreWeave Pushes Prepayments and Short-Term Deals to Fund AI Buildout

CoreWeave is asking customers to pay upfront, borrowing against individual contracts and selling short-term capacity at premium prices — all to fund the infrastructure behind its $132.5 billion backlog.

The neocloud, which rents out computing power for AI workloads, laid out the shift in its third-quarter results. The numbers behind it are big. Revenue hit $3.66 billion, up 168% year on year. But the company also posted a $709 million net loss and paid $846 million in interest expense in a single quarter.

That interest bill explains the financing pivot.

How does CoreWeave plan to fund its expansion?

CEO Michael Intrator said CoreWeave will not use the unsecured debt market at current yields — and doesn't need to. Instead, the company will rely on loans backed by customer contracts to fund most of the infrastructure needed to deliver its $132.5 billion backlog.

The prepayment pipeline is already moving. CoreWeave expects about $9 billion in customer prepayments and deposits over the next three months. More than $3 billion of that ties to capacity it has already built.

In plain terms: customers, not bond investors, are bankrolling the next wave of GPUs.

Why are prices going up?

CoreWeave raised prices by about 25% in July, then another 10% later in the quarter. Intrator said the increases are more than offsetting costs.

The pricing spread between short and long commitments is stark. Short-term capacity is selling at more than $40 million per megawatt annually. Four- and five-year contracts fetch $20 million to $25 million per megawatt.

That gap — roughly double — is why CoreWeave is willing to chase shorter deals despite the added risk. A megawatt sold month-to-month earns far more than one locked up at 2023 prices.

What's the trade-off?

The strategy could lift profits, but it raises CoreWeave's exposure to renewals and pricing. When contracts run short, customers must come back and re-sign — at whatever the market rate is that day. If AI demand cools, that pricing power evaporates fast.

The shift also underscores a broader reality of the AI infrastructure boom: it runs on sustained demand. Every upfront payment CoreWeave collects is a bet by a customer that they'll need this compute — and use it profitably — for years to come.

For now, the bet is paying. Revenue tripled, price hikes stuck, and $9 billion in customer cash is arriving within a quarter. How CoreWeave's model holds up when the short-term contracts come up for renewal will tell us whether the boom has durable footing.