Inside the turmoil at Thinking Machines; sources say Meta discussed buying TML, and CTO Barret Zoph had been in talks since October 2025 about an OpenAI return
Defections, secret conversations, deal talks that fizzled and a battle for control: The turmoil at Thinking Machines Lab …
Context & Ripple Effects
Thinking Machines Lab's control dispute follows a period in which staff departures had already unsettled investors and put future fundraising at risk; the reported leadership and ownership conversations deepen that uncertainty. The earlier wave of staff exits made retention and governance central issues before any strategic transaction was publicly confirmed.
For Meta, the reported outreach fits a broader search for AI-model options: its AI leadership had previously discussed reducing reliance on Llama and using rival models. Those earlier model-strategy discussions give a potential interest in an external lab a clearer strategic context.
First-order effects
- Thinking Machines Lab faces more immediate uncertainty over leadership continuity, control and employee retention as its CTO's reported OpenAI discussions coincide with internal defections.
- Meta's reported acquisition discussions, having not produced a deal, leave it without a confirmed ownership route to TML's team or research while keeping its AI strategy under scrutiny.
Second-order effects
- Investors and prospective hires are likely to weigh the lab's governance risk more heavily, increasing pressure on TML to stabilize leadership before pursuing financing or major partnerships.
- Rival labs, including OpenAI and Meta, can compete more directly for TML talent; later reporting that two founding-team members joined Meta illustrates how personnel movement can substitute for a full acquisition.
Third-order effects
- If frontier labs repeatedly become acquisition targets or talent sources during internal crises, the practical advantage shifts toward incumbents able to absorb teams and finance long research cycles.
- The episode points to governance and key-person retention becoming strategic assets alongside models and compute, though the reported talks alone do not establish a durable consolidation outcome.
The trend: This is one data point in frontier-lab capital concentration, where strategic AI companies increasingly compete through talent, control and access to independent research teams.