Microsoft taps ChatGPT competitor for AI assistant. Is the relationship with OpenAI deteriorating?

Microsoft has started using artificial intelligence from rival OpenAI, the maker of chat bot ChatGPT, in its products. The company announced on September 24 that its AI assistant Copilot will now use the startup Anthropic's models along with those from OpenAI to answer some of its queries.

Anthropic's model is being integrated for the first time into Microsoft 365's Copilot smart assistant, which is capable of analyzing information and creating reports. Users enrolled in the early access program will be able to choose Anthropic's Claude Opus 4.1 AI model as an alternative to the OpenAI model.

Anthropic was founded in 2021 by former top managers of OpenAI. In early September, the company was valued by investors at $183 billion.

Microsoft, which has so far mainly relied on OpenAI models in its products, no longer wants to depend on a single service provider, CNBC explains. At the same time, Microsoft remains OpenAi's largest investor: it has allocated $13 billion to the ChatGPT developer as part of a strategic partnership. OpenAI, in turn, uses Microsoft's Azure cloud capacity.

However, the structure of how Microsoft and OpenAI interact has begun to change, CNBC claims. As recently as last year, Microsoft allowed models from Anthropic and Google to be used in its GitHub developer product Copilot Chat - in addition to OpenAI technologies. Enterprise users can also choose the Claude Sonnet 4 and Claude Opus 4.1 models when building their own Copilot Studio-based agents.

In addition, Microsoft said in late August that it had begun testing its first proprietary AI model, MAI-1-preview, which is fully trained in-house and should improve the capabilities of the Copilot assistant. This could strengthen its own competition with OpenAI, CNBC noted at the time.

Relations between Microsoft and OpenAI are deteriorating and approaching a "boiling point" due to different perceptions of the future of their cooperation in AI, The Wall Street Journal wrote in June. OpenAI would like to reduce Microsoft's control over its products and resources, the newspaper argued. In September, Bloomberg reported, citing OpenAI, that it was closer to resolving "painful" negotiations with Microsoft and turning into a more traditional commercial company.

This month, OpenAI's plans to spend $300 billion on capacity in Oracle's cloud were revealed, and the company also struck a $10 billion deal with chip developer Broadcom. This week, Nvidia said it would invest up to $100 billion to collaborate with OpenAI on a massive data center construction program.

This article was AI-translated and verified by a human editor