Microsoft to integrate Anthropic AI into Office apps, reducing reliance on OpenAI

Microsoft is expanding its artificial intelligence partnerships by striking a deal with Anthropic to use its AI models inside Office 365 applications, according to a report from The Information. The move will see Anthropic’s technology power new features across Word, Excel, Outlook and PowerPoint alongside OpenAI’s, marking a shift away from Microsoft’s earlier strategy of depending solely on the ChatGPT maker.

The decision comes at a time when Microsoft’s relationship with OpenAI appears strained. OpenAI has been building its own infrastructure and exploring ventures such as a potential LinkedIn rival, signalling a more independent direction. Microsoft, meanwhile, is in talks to secure a renewed agreement with OpenAI following its planned for-profit restructuring. However, reports suggest the Anthropic deal is not a negotiation tactic. Executives at Microsoft reportedly believe Anthropic’s latest model, Claude Sonnet 4, outperforms OpenAI’s in specific areas, particularly in generating more polished PowerPoint presentations.

Microsoft has already been diversifying its AI ecosystem. While OpenAI remains the default provider, the company also offers models from Anthropic and xAI through GitHub Copilot. It has also started developing its own systems, unveiling two in-house models named MAI-Voice-1 and MAI-1-preview earlier this year.

OpenAI, for its part, has been working to reduce reliance on Microsoft. The company recently launched a jobs platform to compete with LinkedIn and is preparing to manufacture its own AI chips with Broadcom by 2026. This would allow OpenAI to run training and inference on its own hardware instead of Microsoft’s Azure cloud.