Microsoft Foundry adds Claude. The OpenAI-only era is over.

Microsoft announced on April 27 that Anthropic's Claude models are now available in Microsoft Foundry, alongside the existing OpenAI catalog [Microsoft blog].

── What shipped ──

Foundry now hosts both OpenAI GPT-class models and Anthropic Claude models under the same enterprise security, compliance, and governance framework. Microsoft framed the change as part of "the next phase of the Microsoft-OpenAI partnership."

The framing is corporate-friendly. The substance is that Microsoft is no longer single-vendor on frontier LLMs.

── Why it matters ──

Three years of Microsoft's AI story has been "we are OpenAI's preferred partner." That ends here. Enterprise customers buying through Foundry now get Anthropic without leaving the Azure boundary, which removes the strongest reason to set up direct Anthropic contracts.

The implications:

• OpenAI's leverage with Microsoft drops. When the platform offers competitors at the same compliance level, exclusivity premium collapses. Pricing renegotiations are coming.
• Anthropic's enterprise reach grows materially. Microsoft Foundry's enterprise base is the largest single distribution channel in cloud AI. Anthropic going from "available via direct API" to "available via Foundry" is a step change in reachability.
• Vendor diversity becomes a procurement default. Enterprises that wanted multi-model architectures had to manage multi-cloud or multi-vendor contracts. Now both major frontiers are inside the same Microsoft-managed plane.

── Editor's take ──

The "next phase of the Microsoft-OpenAI partnership" framing is corporate diplomacy. The translation is: Microsoft is hedging. OpenAI's run-rate growth has slowed, Anthropic's has accelerated, and Satya Nadella does not run a $3T market-cap company on a single-vendor bet. The relationship continues. The exclusivity does not.