Two years ago, Satya Nadella paid $650 million to bring Mustafa Suleyman and much of his Inflection AI team to Microsoft. The pitch: Suleyman would make Copilot the AI assistant that 1.5 billion Microsoft users actually wanted to use. That bet has not paid off.
On March 17, Microsoft announced a leadership restructuring that moves Suleyman away from Copilot and into a new role focused on "superintelligence" - essentially, long-horizon model research with deliverables measured in years, not quarters. Jacob Andreou, a former Snap executive who spent eight years scaling that platform, now runs all of Copilot as executive vice president, reporting directly to Nadella.
The Numbers Behind the Move
Copilot has roughly 6 million daily active users as of early March 2026. That puts it behind both ChatGPT and Anthropic's Claude, which sit around 9 million each. More telling: a Recon Analytics survey of over 150,000 U.S. paid AI subscribers found Copilot's market share fell from 18.8% in July 2025 to 11.5% by January 2026. That is a 39% contraction in six months, during a period when the overall paid AI market was growing fast.
Microsoft has Copilot embedded in Windows, Office, Edge, and Bing. It has distribution advantages that OpenAI and Anthropic can only dream about. And it is still losing ground.
A Soft Landing, Not an Exit
Suleyman's new assignment - working on "world-class models for Microsoft over the next five years" - is framed as a promotion toward bigger ambitions. He will sit on a new Copilot Leadership Team alongside Andreou, and says he will "stay directly involved in much of the day-to-day operation." In practice, the person who was hired to own the product no longer owns it.
The restructuring also unifies Copilot's consumer and business divisions under Andreou, collapsing two separate groups that had been operating independently. That fragmentation may explain some of Copilot's identity crisis - it has tried to be a coding assistant, a search replacement, a Windows companion, and an enterprise productivity tool, all at once.
For Microsoft, the $650 million question is whether this is a course correction or an admission that Copilot's distribution-first strategy cannot compensate for a product that users find less compelling than the competition.