CoreWeave, Inc. 8-K
EarningsGuidance deferred to the call
AI-lab and hyperscaler demand drove 113% Y/Y revenue growth; AI-cloud scaling costs pushed GAAP operating income into a loss.
Filed 2026-08-11 16:10 ET
- 2.02 Results of Operations
- 9.01 Financial Statements and Exhibits
Summary
Q2 2026 results, with forward guidance to be provided on the earnings call. Revenue more than doubled as CoreWeave reported accelerating demand from AI labs, hyperscalers and enterprises; it is expanding AI-cloud capacity and adding cross-cloud and agentic AI capabilities. The quarter also showed the cost of scaling: GAAP operating income turned to a loss, while adjusted operating income and adjusted EBITDA margins were lower Y/Y.
Key Figures
| Metric | Q2 2026 | H1 2026 |
|---|---|---|
| Revenue | $2.58B, +113% Y/Y | $4.65B, +112% Y/Y |
| Operating income (loss) | $(49)M, vs $19M | $(193)M, vs $(8)M |
| Adjusted operating income (non-GAAP) | $128M, -36% Y/Y | $149M, -59% Y/Y |
| Net loss | $(626)M, vs $(290)M | $(1.37)B, vs $(605)M |
| Adjusted net loss (non-GAAP) | $(567)M, vs $(130)M | $(1.16)B, vs $(281)M |
| Adjusted EBITDA (non-GAAP) | $1.51B, +100% Y/Y | $2.67B, +96% Y/Y |
| Diluted net loss per share | $(1.14), vs $(0.60) | $(2.53), vs $(1.79) |
| Operating cash flow | $679M, vs $(251)M | $3.66B, vs $(190)M |
- Revenue backlog was ≈$104B at 6/30/2026, excluding more than $25B of net new customer commitments added in early Q3. The company defines backlog as remaining performance obligations plus other estimated future revenue under committed contracts, subject to delivery and service availability.
- Technology and infrastructure expense rose to $1.51B in Q2 from $670M Y/Y. Adjusted operating income was $128M, or 5% of revenue, vs $200M and 16%; adjusted EBITDA margin was 59% vs 62%.
- CoreWeave expanded active power by nearly 500 MW to 1.5 GW and grew contracted power to ≈3.7 GW. It reported customer wins including Bentley Systems, Caterpillar, Grammarly and Isomorphic Labs, and expanded relationships with existing customers including Databricks and Runway ML.
- Product milestones included completing the first industry bring-up and validation of NVIDIA Vera Rubin NVL72, and launching CoreWeave Interconnect, private fiber connections to other cloud platforms beginning with Google Cloud. New agentic AI offerings include ARIA, an AI research and iteration agent, and Sandboxes, secure isolated environments for AI research and evaluation.
- Q2 property and equipment purchases, including capitalized internal-use software, were $6.42B; cash and cash equivalents were $5.52B at quarter-end. The company also reported a $3.1B delayed-draw term loan facility backed by high-performance computing infrastructure, a $1B strategic investment from Jane Street and more than $10B raised through unsecured debt and convertible bonds.
Guidance
Guidance to be given on the earnings call.
Investor Takeaways
- Watch the next backlog update and delivery of the more than $25B in early-Q3 commitments; continued backlog growth alongside conversion into delivered services would support the demand outlook.
- Track active and contracted power against the 1.5 GW and ≈3.7 GW reported for Q2, as capacity expansion is central to serving customer demand.
- Watch the guidance provided on the earnings call and subsequent operating margins; improved adjusted operating income margin from Q2’s 5% would indicate progress toward the company’s stated operating-leverage inflection.
The summary is written by an AI model reading the filing.
Model 005 · prompt v111 · 2026-10-26 11:33:39Z