Mustafa Suleyman co-founded Google DeepMind back when it was just DeepMind, a small London lab. These days he runs Microsoft AI, the part of Microsoft trying to build its own top-tier models instead of relying entirely on OpenAI.
He’s also one of the louder voices on a strange question: what happens when AI seems conscious, whether or not it is?
Who they are
Suleyman grew up in London, left Oxford at 19, and co-founded a youth helpline before getting into tech. In 2010 he co-founded DeepMind with Demis Hassabis. Google bought it in 2014. He left in 2019 and later started Inflection AI with Reid Hoffman, which made a friendly, chatty assistant called Pi.
In March 2024, Microsoft hired him and most of Inflection’s team to run a new division, Microsoft AI. Inflection had raised $1.3 billion in 2023. He also wrote The Coming Wave (2023), a book about how hard powerful AI will be to contain. In March 2026, Microsoft moved its consumer Copilot app under a different executive, and Suleyman’s job narrowed to building Microsoft’s own models and what it calls its superintelligence team.
Why they matter
He’s Microsoft’s backup plan. Microsoft has a big stake in OpenAI, but it doesn’t want to depend on one partner forever. Suleyman’s team released a family of Microsoft-built models at the company’s Build conference in June 2026. One of them, MAI-Thinking-1, can take in about a long book’s worth of text at once.
He’s pushing a “humanist” version of superintelligence. In November 2025, he announced a new team with a pitch that AI should stay clearly under human control. “Humans matter more than AI,” he wrote.
He takes the AI-seems-conscious problem seriously. In August 2025, he wrote an essay warning about “Seemingly Conscious AI,” chatbots convincing enough that people believe they have feelings. His point wasn’t that they do. It was that lots of people will believe it anyway, and that causes real problems. (Tomorrow’s card on AI Consciousness picks this up.)
What they’ve said or done
That August 2025 essay opened with a line that’s hard to forget: “The arrival of Seemingly Conscious AI is inevitable and unwelcome.”
It’s an odd thing for the head of a company’s AI division to say about its own industry. But he’s spent a long time building chatbots meant to feel friendly, so he’d know.

