Google DeepMind co-founder Mustafa Suleyman, who led a heavily criticized unit focusing on practical AI uses, has been placed on leave for unspecified reasons
Context & Ripple Effects
Mustafa Suleyman has always been the political half of DeepMind: a left-leaning activist turned cofounder of the London lab Google acquired for a reported £400M in 2014, and the leader of the unit pushing the research into practical, deployable AI — the very work that drew heavy criticism. His placement on leave, for reasons Google did not specify, removes that applied-AI voice at a moment when DeepMind's real-world deployments were its most contested front.
The leave did not stay mysterious forever, and the corpus shows the arc: sources later attributed the sidelining to alleged bullying of staff, and by early 2022 Suleyman had exited entirely for Greylock Partners. What reads today as an unexplained personnel note was the first step in the founding team's retreat from running the lab.
First-order effects
- Suleyman's sidelining strips DeepMind's applied-AI unit of its champion and public advocate, leaving Demis Hassabis as the only active cofounder in day-to-day management of the lab.
Second-order effects
- The vacuum forced the story into the open: reporting later confirmed the 2019 removal stemmed from alleged bullying of staff, converting a vague leave into a governance-and-culture problem for Google's flagship AI lab.
- The endgame was a permanent loss rather than a return — Suleyman left Google altogether in January 2022 to join Greylock Partners, taking the applied-AI franchise out of DeepMind entirely.
Third-order effects
- The pattern extended to the top: by 2026 Demis Hassabis had drifted from daily CEO duties and stepped down, with Koray Kavukcuoglu installed to lead Google DeepMind under Sundar Pichai — the founding generation fully receding from operations.
- If the sequence holds, founder-led AI labs systematically swap visionary founders for operator leadership once deployment controversies and management-culture issues surface, with investors and boards treating lab culture as a material risk rather than a personnel footnote.
The trend: Founder-led AI labs are being handed over to professional operators, with applied-deployment controversies and staff-culture failures acting as the accelerant.