CoreWeave gave investors a clean read on the artificial intelligence infrastructure boom late Tuesday: demand is still accelerating, but the bill for building enough capacity is enormous.

The Livingston, New Jersey-based cloud company reported second-quarter revenue of $2.575 billion for the period ended June 30, 2026, more than double the $1.212 billion it reported a year earlier. It also reported a $626 million net loss, compared with a $290 million loss in the year-earlier quarter.

The result puts CoreWeave in the middle of a broader market question. AI customers want more computing power, and Wall Street wants to know which companies can turn that demand into durable profit before debt, construction costs and power constraints catch up.

The Numbers

CoreWeave said revenue backlog was approximately $104 billion as of June 30. The company also said that figure does not include more than $25 billion of net new customer commitments added in early third quarter.

Adjusted EBITDA was $1.51 billion, up from $753 million a year earlier, while adjusted operating income fell to $128 million from $200 million. Operating loss was $49 million, compared with operating income of $19 million in the same quarter of 2025.

The company said it expanded active power by nearly 500 megawatts during the quarter to reach 1.5 gigawatts, and grew total contracted power to about 3.7 gigawatts. Those numbers matter because AI cloud revenue depends on more than customer demand. It also depends on chips, power, data center delivery and financing lining up at the same time.

Why Investors Care

CoreWeave is one of the clearest public tests of whether specialized AI cloud providers can scale fast enough to satisfy model developers, large enterprises and hyperscalers. Its customer list in the quarter included AI labs and enterprise names such as Bentley Systems, Caterpillar, Grammarly, Isomorphic Labs and Sunday Robotics, according to the company.

The bullish case is straightforward: if AI workloads keep expanding, CoreWeave's backlog and power commitments could translate into years of revenue growth. The company also pointed to technology milestones, including validation work with Nvidia's Vera Rubin NVL72 platform and new cross-cloud capabilities designed to let customers move workloads across cloud footprints.

The caution is just as clear. CoreWeave's revenue growth comes with large infrastructure spending, large interest expense and execution risk. Interest expense, net, was $640 million in the second quarter. Total liabilities stood at $72.046 billion at June 30, while total assets were $77.070 billion.

The Caveat

A backlog is not the same as cash in the door. CoreWeave defines revenue backlog as remaining performance obligations plus other amounts it estimates will be recognized as revenue in future periods under committed customer contracts, subject to delivery and service-availability requirements.

That caveat is important for readers comparing AI infrastructure stocks. A rising backlog can signal powerful demand, but it still depends on the company delivering capacity on time, keeping customers, managing supplier relationships and funding the buildout without squeezing future returns.

CoreWeave also warned in its forward-looking statement that results could be affected by customer demand, strategic relationships, third-party suppliers and data center partners, capital requirements, tariffs, interest rates and broader geopolitical or macroeconomic conditions.

What To Watch Next

The next test is whether the new commitments CoreWeave described in early third quarter convert into reliable revenue without pushing losses and debt-service costs higher than investors will tolerate.

For the broader AI trade, the report gives a useful signal beyond one stock move. Demand for compute still looks strong. The harder question is who captures enough margin after paying for chips, power, buildings, networking and financing.