Microsoft opens Azure to Anthropic as Nationwide migrates 15M customers to Azure AI

Per runtimewire and Finextra, the strategic headline is Microsoft's shift from OpenAI exclusivity toward multi-model pragmatism: with OpenAI's exclusive arrangement now carrying escape hatches, Microsoft allowed Anthropic to run on Azure — colorfully framed as Microsoft letting 'Anthropic move in while OpenAI was sleeping on the couch.' For enterprise customers, that means Azure increasingly offers a menu (OpenAI, Anthropic and others) rather than a single-vendor bet, mirroring AWS Bedrock's multi-model approach.

The Nationwide deployment grounds the strategy in a concrete enterprise win. The UK financial-services giant ripped out legacy systems and migrated 15 million customers to a new core platform, layering AI across the org: 1,000 premium Copilot licenses in finance, GitHub Copilot for 2,500 engineers, plus Azure OpenAI and Azure AI Foundry. The standout adoption metric — 46% of finance staff using Copilot daily — is the kind of sustained-usage figure enterprises cite to justify AI spend, distinguishing real integration from pilot-stage novelty.

Strategically, this is Microsoft defending its enterprise AI moat on two fronts: broadening model choice on Azure to prevent customers defecting to AWS or GCP for Claude access, and proving deep Copilot penetration in regulated industries. The combination pressures both AWS (on multi-model breadth) and standalone deployments.

Caveats: 'daily use' doesn't quantify productivity gains, and large migrations carry execution risk. The Microsoft-Anthropic-OpenAI triangle also introduces tension — OpenAI is simultaneously OpenAI's biggest partner and a competitor whose model now sits beside Anthropic's on Azure. Watch how the OpenAI relationship evolves, whether more labs land on Azure, and whether Nationwide's usage translates into measurable outcomes.