Thinking Machines Lab Loses Co-Founders to OpenAI
In a surprising leadership shake-up, Thinking Machines Lab—the AI startup founded by former OpenAI CTO Mira Murati—is losing two of its key co-founders, both of whom are returning to OpenAI. The departures of Barret Zoph and Luke Metz mark a significant shift for the well-funded startup just months after it raised $2 billion in seed funding. What’s behind this high-profile talent reversal, and what does it mean for the future of independent AI innovation?
Leadership Exodus Rocks Thinking Machines Lab
On January 14, 2026, Mira Murati announced on X (formerly Twitter) that Barret Zoph, co-founder and former CTO of Thinking Machines Lab, had parted ways with the company. “We have parted ways with Barret,” Murati wrote, quickly adding that Soumith Chintala—longtime AI researcher and PyTorch co-creator—would step into the CTO role. She praised Chintala as “a brilliant and seasoned leader” whose decade-plus contributions to AI made him the ideal successor.
What Murati didn’t mention was that Zoph wasn’t just leaving—he was heading back to OpenAI. Within an hour of her post, Fidji Simo, OpenAI’s CEO of Applications, confirmed that Zoph would rejoin the company alongside fellow Thinking Machines co-founder Luke Metz and former OpenAI researcher Sam Schoenholz. “This has been in the works for several weeks,” Simo noted, signaling a coordinated move rather than a sudden departure.
Why Are Top AI Minds Returning to OpenAI?
The return of Zoph, Metz, and Schoenholz raises questions about the appeal of OpenAI versus newer, independent ventures like Thinking Machines Lab. All three were deeply embedded in OpenAI’s technical DNA before joining Murati’s startup. Zoph served as VP of Research at OpenAI; Metz spent years on its technical staff; Schoenholz led foundational work on neural network stability and generalization.
Their decision to return may reflect shifting dynamics in the AI landscape. As regulatory scrutiny intensifies and competition heats up between Big Tech and agile startups, the resources, infrastructure, and global reach of established players like OpenAI could be proving irresistible—even to those who once sought independence.
Moreover, OpenAI’s aggressive product roadmap under CEO Sam Altman and applications chief Fidji Simo suggests a renewed focus on deploying cutting-edge models at scale. For engineers and researchers eager to see their work impact millions of users, that momentum may outweigh the allure of building something from scratch.
A New CTO Steps In Amid Uncertainty
While the departures are undoubtedly a blow, Murati’s swift appointment of Soumith Chintala as CTO signals resilience. Chintala isn’t just a respected name in machine learning—he’s a battle-tested builder. As co-creator of PyTorch, one of the most widely used deep learning frameworks in the world, he brings both technical depth and community credibility.
His promotion also underscores Thinking Machines Lab’s commitment to open, developer-friendly AI—a philosophy that aligns with Murati’s public stance on responsible innovation. With Chintala at the helm of engineering, the startup may double down on its differentiation: building advanced AI systems with transparency, safety, and accessibility baked in from day one.
Still, the loss of two co-founders so early in the company’s life is a red flag investors and partners will be watching closely. Can a startup valued at $12 billion maintain its trajectory when core architects walk away less than a year after launch?
The $2 Billion Question: What’s Next for Thinking Machines Lab?
Thinking Machines Lab emerged in mid-2025 with fanfare, backed by a staggering $2 billion seed round led by Andreessen Horowitz and joined by heavyweights like Nvidia, AMD, and Jane Street. The valuation—$12 billion out of the gate—reflected sky-high expectations for Murati’s vision: an AI lab focused on long-term, foundational research without the pressure of quarterly earnings or product deadlines.
But the AI industry moves fast. In 2026, the race isn’t just about who builds the smartest model—it’s about who can deploy it safely, ethically, and profitably. OpenAI, Google DeepMind, and Anthropic are all accelerating their commercial strategies, while regulators in the U.S. and EU push for stricter oversight.
In this climate, even well-funded startups face existential choices. Murati’s team must now prove that Thinking Machines Lab isn’t just a haven for ex-OpenAI talent—but a durable, independent force capable of shaping the next era of artificial intelligence.
Murati Stays the Course—For Now
Despite the upheaval, Mira Murati remains at the helm as CEO. Her leadership has been defined by a rare blend of technical rigor and ethical clarity—a combination that earned her respect across Silicon Valley during her tenure at OpenAI. If anyone can navigate this turbulence, it’s her.
Yet the optics are challenging. Losing co-founders to your former employer—especially one as dominant as OpenAI—can erode confidence among employees, investors, and potential partners. Murati’s next moves will be critical: Will she pivot the company’s strategy? Accelerate product timelines? Or double down on pure research and let the market catch up?
One thing is clear: the AI gold rush is entering a new phase. It’s no longer enough to assemble a dream team—you have to keep them.
Talent Wars Heat Up in AI
This episode reflects a broader trend in the AI sector: intense competition for top talent. As companies vie to lead the next wave of intelligent systems, poaching, counter-offers, and strategic returns are becoming commonplace. What makes this case unusual is the speed and symmetry—three senior figures leaving a hot startup to rejoin the very company they left to escape.
It also highlights the gravitational pull of OpenAI. Despite internal controversies and leadership drama in recent years, the company remains a magnet for elite AI researchers—thanks to its data advantages, compute resources, and real-world deployment scale.
For aspiring AI entrepreneurs, the message is sobering: building a compelling alternative to tech giants requires more than capital and charisma. It demands a mission so compelling that even the best minds choose to stay—even when the door back to OpenAI swings open.
As 2026 unfolds, all eyes will be on Thinking Machines Lab. Can it weather this storm and emerge as a true counterweight to the AI establishment? Or will it become another cautionary tale of how hard it is to leave the giants behind?
