Microsoft expands AI roster with Anthropic and xAI integrations, looking beyond OpenAI alliance

Microsoft is continuing to expand its fleet of AI partners, hedging its bets beyond its high-profile partnership with OpenAI — a change that’s coming into sharper focus this week at the company’s Build developer conference. On stage this morning in downtown Seattle, the company announced that Anthropic’s Claude Code will be available as part of Microsoft’s newly unveiled GitHub Copilot autonomous coding agent, in addition to OpenAI’s Codex agent. This came a day after Microsoft CEO Satya Nadella said the company would offer Grok 3 and Grok 3 Mini, developed by Elon Musk’s xAI, via Microsoft’s Azure AI Foundry AI… Read More

May 21, 2025 - 03:30

Jay Parikh, Microsoft executive vice president and head of the company’s CoreAI engineering division, speaks at Build 2025 in Seattle, with the Anthropic and OpenAI logos highlighting Microsoft’s push to support multiple AI models. (GeekWire Photo / Todd Bishop)

Microsoft is continuing to expand its fleet of AI partners, hedging its bets beyond its high-profile partnership with OpenAI — a change that’s coming into sharper focus this week at the company’s Build developer conference.

On stage this morning in downtown Seattle, the company announced that Anthropic’s Claude Code will be available as part of Microsoft’s newly unveiled GitHub Copilot autonomous coding agent, in addition to OpenAI’s Codex agent.

This came a day after Microsoft CEO Satya Nadella said the company would offer Grok 3 and Grok 3 Mini, developed by Elon Musk’s xAI, via Microsoft’s Azure AI Foundry AI platform.

“Openness and choice — that’s what all developers want, and that’s what we’re going to deliver,” said Jay Parikh, Microsoft executive vice president and head of the new CoreAI engineering division.

All of this comes at a time of transition for Microsoft’s relationship with OpenAI, its most prominent AI partner, in which the Redmond company has invested more than $13 billion.

Microsoft and OpenAI are reportedly renegotiating their multibillion-dollar partnership, according to the Financial Times, as OpenAI restructures into a new public benefit corporation.

A recent Businessweek deep dive on Microsoft described the alliance as evolving from a close collaboration into a more complicated dynamic, with Microsoft increasingly pursuing its own models and alternative partnerships to reduce long-term dependence on OpenAI.

Microsoft acquired the GitHub software development platform in 2018. OpenAI’s Codex agent provided the early foundation for the GitHub Copilot pair programming tool, and remains a core option for developers, but Microsoft last year started expanding GitHub Copilot with other models, including Anthropic’s Claude.

That broader approach, incorporating different AI technologies, is now continuing with GitHub Copilot’s agentic capabilities, announced this week.

Highlighting its role as both Microsoft competitor and partner, OpenAI last week announced a new research preview of a standalone version of Codex, which is designed to take on larger, multi-step programming tasks more autonomously.

After showing up at Build in person last year, OpenAI CEO Sam Altman appeared virtually at the conference on Monday, joining Nadella to discuss the evolution of AI tools and coding agents.

Altman called the current shift to autonomous agents “one of the biggest changes in programming I’ve ever seen,” saying tools like Codex are becoming more like virtual teammates — able to take on complex tasks, respond in parallel, and work deeply within GitHub environments.

He predicted rapid improvements in model reliability, tool use, and integration, saying that developers should expect models to become “more similar to human teammates” over time.

OpenAI still relies heavily on Microsoft’s technology. Wrapping up the Build technical keynote on Tuesday morning, Scott Guthrie, executive vice president of Microsoft’s Cloud + AI Group, used OpenAI’s ChatGPT as a case study to highlight the scale and capabilities of Microsoft’s Azure infrastructure.

He said the cost of running large models like GPT-4 has dropped by 93% in the past two years, enabling a wave of new AI applications and agents.

ChatGPT, which he called the “canonical app” of this AI era, runs entirely on Azure — using services such as Azure VMs, Kubernetes, Cosmos DB, Postgres, and Azure Storage.

Microsoft is competing against other major cloud platforms — including Anthropic’s partner and investor Amazon — in an effort to expand AI capabilities for developers and businesses.

Google, which has also invested in Anthropic, kicked off its I/O developer conference this morning in California, announcing its own new AI and cloud tools.

Microsoft CTO Kevin Scott used his remarks to analysts and reporters in advance of Build to call for a more open and connected future for AI. He said the industry should learn from the early days of the internet, when closed systems nearly held back the web’s potential.

Scott said AI agents will only reach their full potential if they can “talk to everything in the world.”