CoreWeave expands full-stack AI cloud push as inference demand grows

CoreWeave Inc., the AI-native cloud provider, is expanding its full-stack AI cloud offerings as demand for inference grows. The company recently validated Nvidia's Vera Rubin NVL72, marking the industry's first bring-up and validation on CoreWeave Cloud. Enterprise interest in AI is shifting from training to inference, prompting neoclouds like CoreWeave to provide tailored cloud infrastructure.

86% of enterprises prioritize data unification over compute, emphasizing that AI performance relies on more than just accelerators. CoreWeave's support for Nvidia's comprehensive, unified platform, designed for agentic AI, highlights this shift in AI infrastructure focus.

According to Paul Nashawaty, principal analyst for theCUBE Research, "CoreWeave's evolution highlights a fundamental shift in how enterprises should think about AI infrastructure." The company's expertise across the entire infrastructure stack, including scalability, observability, security, governance, and cost management, positions it to differentiate itself in the evolving AI marketplace.

As enterprises struggle to operationalize AI, with 30% facing operational readiness gaps and 88% of AI pilots failing to reach production, CoreWeave aims to narrow this gap through its purpose-built stack. Jean English, CoreWeave's chief marketing officer, emphasizes the platform's cost advantages, boasting one-tenth the cost per million tokens compared to Nvidia's previous releases.

The company's collaboration with Nvidia is just the beginning, with Chief Technology Officer Peter Salanki envisioning a future where inference is disaggregated into specialized pipeline stages, with smaller models handling queries and larger models taking over for more complex tasks.