Anthropic's Claude Hits Azure Foundry with NVIDIA GB300

Anthropic made its Claude models generally available in Microsoft Foundry on Tuesday, adding a third major cloud deployment for the AI startup behind the Opus and Haiku models. Foundry is Microsoft's hub for building, deploying, and managing AI applications on Azure. The rollout includes the latest Claude versions – Opus 4.8 and Haiku 4.5 – and support for NVIDIA's GB300 GPUs.

The move ties Microsoft deeper into the enterprise AI arms race. AWS already offers Claude via Amazon Bedrock; Google Cloud has it on Vertex AI. Azure clients now get a native route, which means one less cloud-switch for Microsoft-heavy enterprises. The GB300 GPU inclusion also signals that NVIDIA's next-generation hardware is finding a home in production inference workloads, not just training.

Microsoft shares, at $368.50, were down 1.2% on Tuesday. NVIDIA was up 1.2% at $194.90. The divergent moves reflect a market that sees the cloud platform getting stickier but remains skeptical about near-term AI monetization. AlphaScala's Alpha Score, which blends technical and fundamental signals, rates MSFT at 42 out of 100 (Mixed) and NVDA at 62 (Moderate).

Claude's expansion into Foundry gives Microsoft another arrow in its enterprise AI quiver – one that competes directly with its own OpenAI partnership. For customers who prefer Claude's safety stance or its longer context window, the option now lives inside their existing Azure subscription. The GB300 GPU availability suggests NVIDIA is betting this deployment pattern will scale, even if the revenue impact takes quarters to appear.