CoreWeave announced that it will be among the first cloud providers to deploy NVIDIA’s Rubin technology in the second half of 2026. The integration is designed to accelerate support for agentic AI, advanced reasoning and large-scale inference workloads across the company’s GPU-optimized infrastructure. Rubin’s architecture promises up to 3× faster inference performance compared with previous-generation accelerators, enabling CoreWeave to offer differentiated service levels for enterprise customers tackling real-time decisioning, autonomous agent frameworks and data-intensive models. CoreWeave has already secured capacity reservations on over 10,000 next-generation GPUs and expects the Rubin rollout to drive usage growth of 40% year-over-year by late 2026, as enterprises migrate proof-of-concepts into production environments.

After maintaining an Underperform rating since mid-2025, D.A. Davidson analyst Gil Luria upgraded CoreWeave to Neutral, citing accelerating customer adoption of its AI cloud services and strong balance-sheet positioning ahead of the Rubin deployment. Luria highlighted that CoreWeave ended 2025 with $350 million in annualized recurring revenue, a 120% increase over the prior year, and maintained an unlevered free cash flow margin above 25%. The upgrade reflects confidence in the company’s ability to convert enterprise pilot projects into multi-year contracts, and the analyst adjusted the 12-month consensus price target upward by more than 35% based on revised margin assumptions and capacity expansion plans.