CoreWeave inks multiyear cloud deal with Anthropic

2026-04-10 · Source: AI – SiliconANGLE · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

CoreWeave Inc. secured a multiyear contract to provide cloud infrastructure to Anthropic PBC, causing CoreWeave's shares to close 11% higher. This infrastructure will support both training and inference workloads for Anthropic's Claude models, utilizing a "variety" of Nvidia Corp. chips in U.S. data centers, including the advanced Blackwell Ultra. The Blackwell Ultra features two four-nanometer dies, NV-HBI interconnect for 10 terabits per second data streaming, warp-synchronous memory, and Special Function Units for transcendental number operations. This deal follows CoreWeave's expanded agreement with Meta Platforms Inc. for initial deployments of the Blackwell Ultra's successor, Vera Rubin. CoreWeave operates over 43 data centers with 850 megawatts capacity, offering single-tenant Dedicated Access AZs and custom elements for workload prediction and cooling optimization. The company claims to serve 9 of the top 10 AI model providers, including OpenAI, which previously committed to $22.4 billion in infrastructure.

Key takeaway

For CTOs and investors evaluating AI infrastructure providers, CoreWeave's recent multiyear contract with Anthropic, following a $22.4 billion deal with OpenAI, signals its strong market position and capacity to support leading AI model development. You should consider CoreWeave for high-performance, dedicated AI cloud infrastructure, especially given its early access to advanced Nvidia GPU technologies like Blackwell Ultra and Vera Rubin, which are critical for scaling large language model training and inference.

Key insights

CoreWeave's strategic partnerships and advanced Nvidia-powered infrastructure are key to its rapid growth in AI cloud services.

Principles

• Dedicated infrastructure enhances AI workload performance.
• Specialized hardware accelerates AI model development.
• Scalable cloud capacity is critical for AI innovators.

Method

CoreWeave deploys single-tenant Dedicated Access AZs with custom cooling and power distribution to optimize for AI workload spikes, leveraging advanced Nvidia GPUs like Blackwell Ultra and its successor, Vera Rubin.

In practice

• Utilize single-tenant cloud for sensitive AI workloads.
• Prioritize GPUs with specialized AI acceleration features.
• Consider cloud providers with custom data center optimizations.

Topics

• CoreWeave
• Anthropic
• Cloud Infrastructure
• NVIDIA Blackwell Ultra
• AI Model Deployment

Best for: CTO, Investor, Director of AI/ML, VP of Engineering/Data, Tech Journalist

Related on AIssential

• Perplexity, CoreWeave Deal Boosts Inferencing — The AI market is shifting focus from training to inference, driving specialized cloud partnerships for real-time AI applications.
• Jane Street Invests $1 Billion in CoreWeave — Jane Street invested $1B in CoreWeave and committed $6B for AI cloud access, including Nvidia's Vera Rubin technology.
• Claude Users FINALLY Get More Usage — Anthropic expanded Claude usage and secured major compute deals with SpaceX and Google Cloud.
• CoreWeave, Meta Strike $21 Billion for AI Computing — The AI sector sees massive compute infrastructure investments, strategic model development, and complex financing, alongside evolving geopolitical and market dynamics.
• Inside the $41B AI Cloud Challenging Big Tech | CoreWeave SVP — Specialized AI cloud providers can outperform general public clouds for GPU-intensive AI training workloads.
• Jane Street commits $6bn to CoreWeave AI cloud partnership — Strategic partnerships between financial firms and AI cloud providers drive advanced machine learning in trading.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI – SiliconANGLE.