Mira Murati’s Thinking Machines Lab Loses Key Leaders as They Return to OpenAI

Thinking Machines Lab — the startup launched by former OpenAI CTO Mira Murati — is hemorrhaging key leadership as co-founders abandon ship to return to their former employer.

Two co-founders, Barret Zoph and Luke Metz, are heading back to OpenAI, alongside Sam Schoenholz, another former OpenAI staffer who had joined the startup.

This reversal comes just six months after the company secured a record-breaking $2 billion funding round that valued the startup at $12 billion. The timing couldn’t be more devastating—especially considering this marks the second major departure from Thinking Machines Lab in recent months, following co-founder Andrew Tulloch’s exit to Meta last October.

Zoph was actually fired for “unethical conduct” after sources revealed he had shared confidential company information with competitors. The drama unfolded rapidly—Zoph told Murati he was considering leaving, then found himself terminated.

See you stars

The departures represent a massive brain drain for what was supposed to be AI’s next big thing. Zoph previously served as VP of research at OpenAI and spent six years at Google as a research scientist. Metz had spent years on OpenAI’s technical staff. Their combined expertise was considered crucial to the startup’s mission of building multimodal AI that works with how humans naturally interact with the world.

The seed round from six months ago attracted top-tier investors including Andreessen Horowitz, Accel, Nvidia, AMD, and Jane Street—making this talent exodus particularly devastating for those betting billions on Murati’s vision.

The departures were announced in a rather dramatic fashion, with Murati revealing the split and OpenAI’s CEO of applications Fidji Simo later providing full details. Simo stated that the moves had been in planning for weeks, suggesting this wasn’t a spontaneous decision.

Soumith Chintala will become the new CTO of Thinking Machines Lab, but the leadership upheaval raises serious questions about the startup’s ability to execute on its ambitious vision.

The situation is fluid as media reports suggest that OpenAI is also planning to bring over more researchers from Thinking Machines.

What this talent reversal means for AI’s future

This exodus reveals deeper instabilities in the AI startup ecosystem. Thinking Machines Lab was supposed to represent a new model for ethical AI development, emphasizing transparency and open science principles. The company had assembled a team of approximately 30 top-tier experts from leading organizations such as OpenAI, Meta, and Mistral.

The startup was building “agentic AI” systems that can actually take autonomous action, representing a significant leap beyond current conversational AI models. This technology was expected to transform everything from drug discovery to software development.

But the broader implications extend far beyond one startup’s struggles. This talent reversal suggests that despite massive funding rounds and ambitious visions, the gravitational pull of established AI giants like OpenAI remains incredibly strong. Even more telling: this is happening while OpenAI loses $83 million a day and enterprise adoption has dropped from 14% to 12% based on U.S. Census Bureau data.

The timing is particularly significant given the volatile state of the AI industry, where 95% of pilots fail to reach production per McKinsey research, and three firms control 60% of Gen AI funding, all burning billions.

The talent war

For investors who poured billions into Thinking Machines Lab’s vision, this development serves as a sobering reminder that in the AI world, human capital remains the most valuable and volatile asset.

The company’s first product, which was expected to include a significant open source component, now faces an uncertain timeline as the startup rebuilds its leadership team.

OpenAI has dropped news of a massive partnership that promises to deliver the fastest AI responses ever seen.