Microsoft Launches Its Own AI Models to Rival OpenAI

Microsoft has spent years and many billions of dollars bankrolling the two biggest names in AI. This week it started competing with them.

At its annual Build developer conference, the company unveiled a family of seven AI models built entirely in-house, branded MAI. The message is hard to miss: Microsoft would rather build its own frontier models than keep renting them.

A frontier model is one of the large, general-purpose AI systems at the cutting edge of what the technology can do, the kind that power chatbots and coding assistants. Until now, Microsoft mostly got those from OpenAI, the maker of ChatGPT, which it has backed with around $13 billion (€11.8bn). It has also put up to $5 billion (€4.5bn) into Anthropic, the lab behind the Claude chatbot. Both companies' models run on Microsoft's Azure cloud.