CoreWeave Q2 2026 results showed AI demand driving pricing gains and capacity growth.

CoreWeave's Q2 FY 2026 earnings show demand for purpose-built AI cloud capacity continuing to exceed available supply, with backlog, active power, and managed inference adoption all moving higher. The quarter also shows CoreWeave shifting from pure infrastructure scale-out toward a broader AI platform model that ties training, inference, observability, and agent development together.

CoreWeave (NASDAQ: CRWV) reported Q2 FY 2026 revenue of $2.58 billion, up 112% year over year, compared with Wall Street consensus of $2.56 billion. Adjusted operating income was $128 million, down from $200 million a year ago, with an adjusted operating income margin of 5% (versus 16% in Q2 FY 2025). Adjusted net loss was $567 million, compared with $130 million a year ago. Diluted net loss per share was $1.14, compared with $0.60 a year ago.

"CoreWeave reached an important inflection point this quarter as our scale began to translate into expanding operating leverage," said Michael Intrator, co-founder, chairman, and chief executive officer of CoreWeave. "Customer demand is accelerating, as enterprise adoption broadens and we continue to deepen our technology platform. CoreWeave is built on the conviction that AI is foundational to every industry and that realizing its full potential requires a purpose-built platform. This quarter reinforced that conviction."

CoreWeave's Q2 results show a company still absorbing heavy infrastructure investment, but with clearer evidence that scale is starting to improve operating performance. The central market signal is not only revenue growth, but also the pricing power tied to constrained AI cloud supply. CoreWeave is expanding beyond GPU rentals into managed inference, agentic AI tooling, observability, and cross-cloud services, as AI demand shifts toward continuous training, inference, evaluation, and improvement cycles. This underscores the view that AI infrastructure providers with power access, supply-chain depth, and software attachment can capture more durable economics than capacity-only providers.

CoreWeave ended Q2 FY 2026 with 1.5 gigawatts of active power after adding nearly 500 megawatts during the quarter. Contracted power reached 3.7 gigawatts at quarter-end and 4.2 gigawatts as of the earnings call. Near-term capacity remains effectively sold out, with demand from multiple customers for each GPU brought online. New Q2 FY 2026 customer contracts are expected to carry contribution margins 5 to 10 percentage points above recent quarters. CoreWeave raised pricing across SKUs by approximately 25% in July while passing through component cost increases. Scarce energized capacity gives CoreWeave more control over customer selection, pricing, and contract structure.

Managed inference is becoming a strategic extension of CoreWeave's infrastructure base rather than a separate product motion. Booked ARR for Managed Inference grew from $1 million to more than $100 million within a few months of launch. The offering lets CoreWeave monetize tokens while giving customers more flexible consumption choices across serverless and dedicated deployments. Grammarly and You.com are using the platform for production traffic, including AI coding agents, fine-tuned models, and open-weight model deployments. Older GPU fleets also create a useful supply source for inference as original contracts expire, providing CoreWeave a higher-margin path to deepen customer spend after the initial infrastructure sale.

Q2 FY 2026 customer activity showed demand expanding across AI labs, enterprises, financial services, industrial systems, life sciences, and public sector work. New customer wins included Bentley Systems, Caterpillar, Grammarly, Isomorphic Labs, and Sunday Robotics, while expanded relationships included Cognition, Databricks, Hudson River Trading, Periodic Labs, Rescale, and Runway ML. Caterpillar is using CoreWeave's AI cloud infrastructure to support physical AI training and inference for autonomous construction equipment. Flow Traders and IMC extend CoreWeave's presence in systematic trading, where low-latency and high-throughput AI workloads are becoming more common. The Leidos collaboration adds federal demand tied to defense, national security, and intelligence missions. Broader vertical adoption reduces CoreWeave's dependence on a narrow set of AI-native customers over time.

CoreWeave guided Q3 FY 2026 revenue to $3.45 billion to $3.60 billion and adjusted operating income to $200 million to $260 million. Q3 FY 2026 capital expenditures are expected to be $11.5 billion to $13.5 billion, with adjusted operating margin expected to reach the low teens in Q4 FY 2026. For FY 2026, CoreWeave raised revenue guidance to $12.4 billion to $13.2 billion (prior guidance: $12.0 billion to $13.0 billion) and adjusted operating income guidance to $960 million to $1.15 billion (prior: $900 million to $1.1 billion). FY 2026 capital expenditures are expected to be $35 billion to $39 billion, while end-of-year annualized run-rate revenue is expected to reach $18.5 billion to $19.5 billion (prior: $18.0 billion to $19.0 billion).

The raised outlook reflects continued demand for AI infrastructure, though the scale of CoreWeave's capacity expansion keeps execution and capital intensity central to the story. Managed Inference, broader enterprise adoption, and improving economics on new customer contracts provide additional paths to monetize the infrastructure being deployed. At the same time, the company must bring contracted power online quickly enough to support backlog conversion and improve operating leverage. If CoreWeave can translate its expanding power footprint into highly utilized capacity while increasing inference and software attachment, the current infrastructure buildout could support a more diversified and scalable AI cloud model.