CoreWeave raises revenue guidance as AI demand builds $104B backlog
CoreWeave reported second-quarter revenue of $2.58 billion, a 112% year-over-year increase, while raising its full-year revenue guidance to a range of $12.4 billion to $13.2 billion. The AI data center provider is rapidly scaling its infrastructure capacity, targeting 1.85 gigawatts by year-end with increased capital expenditure plans of up to $39 billion.
Key Takeaways
- Revenue backlog reached $104 billion, with an additional $25 billion in new customer commitments secured early in Q3.
- Capital expenditure guidance increased to $35 billion–$39 billion to fund data center capacity targeting 1.85 gigawatts by year-end.
- Net losses widened to $626 million from $290 million a year ago, primarily driven by costs associated with $35 billion in total debt.
- Management implemented a 25% price increase across all SKUs in July, citing the ability to pass through rising component costs to customers.
Why It Matters
CoreWeave’s performance confirms that the market for high-density AI compute remains severely supply-constrained, giving specialized providers significant pricing power over hyperscalers. For the streaming ecosystem, this indicates that the infrastructure costs for AI-driven personalization and video encoding will remain elevated as providers compete for GPU access. The surge in capital spending highlights a growing divide between traditional cloud providers and 'neoclouds' that prioritize immediate hardware availability over long-term profitability. Industry observers should watch for potential delivery delays as CoreWeave navigates regional regulatory moratoriums on new data center construction in states like New York.
Additional Context
CoreWeave’s rapid expansion is occurring alongside a massive shift in the U.S. power landscape. Per Goldman Sachs (May 2026), data center electricity demand is projected to double to 66 gigawatts by 2027, potentially accounting for 8.5% of total U.S. peak summer power demand. This tightening energy market has forced infrastructure providers to compete for grid connections as aggressively as they do for silicon. While CoreWeave targets 8 gigawatts of capacity by 2030, analysts at BCG have warned of a potential 50-to-80 gigawatt shortfall in total U.S. capacity by that date, suggesting that power availability—rather than chip supply—may become the primary bottleneck for AI scaling. Competition is also diversifying beyond traditional cloud giants. Per Reuters (August 2026), SpaceX has entered the market by selling excess compute capacity, leveraging its own exclusive hardware partnership with Nvidia to secure priority chip allocations. This move by SpaceX, combined with Meta’s reported exploration of a dedicated cloud unit, places new pressure on CoreWeave’s 'neocloud' model. Despite these entrants, CoreWeave’s July price increases of approximately 25% across its SKUs, as reported by Seeking Alpha (August 2026), indicate that near-term demand from entities like Anthropic and Jane Street Group remains high enough to absorb both rising hardware costs and increased competition.
Read full article at siliconangle.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source