Claude Meets Blackwell Ultra: Anthropic’s Models Now Run on NVIDIA GB300 in Azure

Anthropic’s Claude models are now generally available on Microsoft Azure via Microsoft Foundry, running on NVIDIA’s new GB300 Blackwell Ultra GPUs. This partnership gives Azure‑native enterprises a high‑performance, low‑latency platform for building autonomous and domain‑specific AI agents. The deployment leverages Azure’s managed services while exposing the cutting‑edge GPU architecture for scalable inference, though it currently supports only Claude models and requires an Azure subscription with GPU access.

⚡ Key Takeaways

• Claude models now run on NVIDIA GB300 Blackwell Ultra GPUs in Azure.
• Deployment is managed through Microsoft Foundry’s Azure AI Studio.
• Requires Azure subscription and GPU availability, limiting immediate access to organizations without GB300 capacity.
• Deploy by selecting the GB300 GPU type in the Azure AI Studio deployment settings.
• Currently limited to Claude; other LLMs are not yet supported on this GPU stack.
• WhyItMatters: Engineers shipping production AI can now leverage Azure’s managed services combined with NVIDIA’s latest GPU architecture to achieve faster inference and lower latency for Claude‑based agents, directly impacting throughput and cost efficiency.
• TechnicalLevel

Engineers shipping production AI can now leverage Azure’s managed services combined with NVIDIA’s latest GPU architecture to achieve faster inference and lower latency for Claude‑based agents, directly impacting throughput and cost efficiency. TechnicalLevel

Want the full story? Read the original article.