Satya Nadella, who revived Microsoft by betting early on cloud computing, now faces a fresh test as he tries to keep the company central to the AI boom. Microsoft's once-exclusive partnership with OpenAI has loosened: the software giant is no longer the lab's exclusive cloud provider, and its license on OpenAI's intellectual property became non-exclusive this year.
A key point of uncertainty is Mustafa Suleyman, the AI researcher Nadella hired in 2024 to lead AI initiatives. Suleyman's team has shipped models for coding, image generation, and speech-to-text, and he recently boasted on X that a new transcription model was "the most accurate real time transcription model in the world." Yet on independent leaderboards, these models have yet to match the raw capability of Anthropic's or OpenAI's offerings. Microsoft executives argue that keeping model development and application teams together has advantages: the group building Teams and other Office apps has seen better performance from Suleyman's transcription models at a lower price, said Executive Vice President Charles Lamanna.
In March, Microsoft merged the engineering teams for its commercial and consumer Copilot assistants after sluggish adoption. Jacob Andreou now leads the Copilot experience, while Suleyman focuses on model development. Nadella, for his part, has played down the importance of a single world-beating model. In June he wrote that the priority must be building a "frontier ecosystem," not just a frontier model, so value flows broadly across every company, industry, and country. At a September event, he stood before a slide declaring Copilot could connect to "every model," with logos from Anthropic, OpenAI, SpaceX, and Microsoft AI.