CoreWeave shares surged 21% after the AI cloud provider posted its strongest earnings report yet, with second-quarter revenue more than doubling to $2.58 billion from $1.21 billion a year earlier and edging past Wall Street's $2.56 billion estimate. The company's adjusted loss per share came in at $1.03, narrower than the $1.20 loss analysts had modeled, even as its net loss widened to $626 million from $290 million a year ago on rising interest costs.
The headline figure driving investor enthusiasm was backlog: in its own earnings release, CoreWeave said revenue backlog reached roughly $104 billion as of June 30, with more than $25 billion in additional net-new customer commitments signed in the early weeks of the third quarter, pushing the combined figure toward $129 billion.
Scaling infrastructure to match demand
To support that pipeline, CoreWeave raised its year-end target for active power capacity to more than 1.85 gigawatts and guided to $35 billion to $39 billion in capital expenditures for the year. Technology and infrastructure expenses climbed to $1.51 billion in the quarter from $670 million a year earlier as the buildout accelerated. Adjusted EBITDA came in at $1.51 billion, a 59% margin, down slightly from 62% a year ago as the company absorbed the cost of scaling.
Marquee customer deals underpin the backlog
Much of the backlog growth traces to a handful of large enterprise commitments. CoreWeave signed a $21 billion agreement to supply AI cloud capacity to Meta through 2032, layered on top of a prior $14 billion commitment, and reached a separate multi-year deal to provide compute for Anthropic's Claude models. Jane Street also made a $1 billion strategic investment in the company following an expanded commercial relationship struck earlier in the year. For 2026, CoreWeave now guides to $12.4 billion to $13.2 billion in full-year revenue and $960 million to $1.15 billion in adjusted operating income.
Related: Hyperscalers Commit Record $2.6 Trillion in AI Infrastructure Spending
The results land amid a broader surge in hyperscaler AI infrastructure spending industry-wide, and CoreWeave's growing backlog — now nearly 50 times its quarterly revenue — signals that demand for dedicated AI compute capacity is still outpacing what providers can build, even as the widening losses underscore how capital-intensive that buildout remains.
