The artificial intelligence industry, already known for its rapid innovation and fierce competition for talent, is witnessing another high-profile reshuffling—this time involving one of the most closely watched AI startups founded by former OpenAI leadership. Thinking Machines Lab, the ambitious venture launched by ex-OpenAI CTO Mira Murati, is undergoing a significant internal transition as two of its co-founders and another senior researcher depart the company and return to OpenAI.
The departures, announced publicly through a series of social media posts, underscore both the volatility of the AI startup ecosystem and the gravitational pull OpenAI continues to exert on elite research talent. They also raise questions about the internal dynamics of Thinking Machines, a company that has raised unprecedented funding for a startup at such an early stage and positioned itself as a future leader in frontier AI development.
A Sudden Announcement — and a Swift Countermove
On Wednesday, Mira Murati took to social media platform X to announce that Barret Zoph, co-founder and Chief Technology Officer of Thinking Machines Lab, was leaving the company. The announcement was notably brief and formal in tone.
“We have parted ways with Barret,” Murati wrote. She went on to reveal that Soumith Chintala—an influential AI researcher best known for his work on PyTorch—would step into the role of CTO. Murati praised Chintala as “a brilliant and seasoned leader” with over a decade of impactful contributions to the field of artificial intelligence, emphasizing his role as a key contributor within Thinking Machines since its inception.
The message was clear in terms of leadership succession, but sparse in detail regarding the circumstances of Zoph’s exit. Murati’s post made no reference to additional departures, nor did it elaborate on the reasons behind the separation.
Less than an hour later—just 58 minutes after Murati’s post—OpenAI effectively reframed the narrative.
Fidji Simo, OpenAI’s CEO of Applications, announced on X that Barret Zoph would be returning to OpenAI, alongside Luke Metz and Sam Schoenholz, both of whom were also affiliated with Thinking Machines.
“Excited to welcome Barret Zoph, Luke Metz, and Sam Schoenholz back to OpenAI!” Simo wrote. “This has been in the works for several weeks, and we’re thrilled to have them join the team.”
The timing and wording of the announcement suggested that these moves were not abrupt, but rather the culmination of discussions that had been ongoing behind the scenes.
The Co-Founders at the Center of the Shift
Barret Zoph’s departure is particularly consequential. Prior to co-founding Thinking Machines Lab, Zoph served as Vice President of Research at OpenAI, where he played a key role in advancing the organization’s AI research agenda. Before that, he spent six years at Google as a research scientist, building a reputation as a leading figure in machine learning research.
Luke Metz, another co-founder of Thinking Machines, also brings deep OpenAI credentials. Metz previously worked on OpenAI’s technical staff for several years and was closely involved in advanced research initiatives. His decision to return to OpenAI alongside Zoph signals a significant realignment of talent.
Sam Schoenholz, whose LinkedIn profile until recently listed him as a member of Thinking Machines, likewise has deep roots at OpenAI. His inclusion in OpenAI’s announcement confirms that the talent exodus extends beyond just the company’s co-founders.
Together, the departures represent the loss of a substantial portion of Thinking Machines’ founding technical leadership—a rare and notable development for a startup that is less than a year old.
A Startup Born from OpenAI’s Upper Ranks
Thinking Machines Lab was founded by Mira Murati after her departure from OpenAI in September 2024, where she had served as Chief Technology Officer. Murati was widely regarded as one of the most influential leaders within OpenAI, overseeing critical aspects of product development and research execution during a period of explosive growth for the organization.
Following her exit, Murati joined forces with Zoph and Metz to launch Thinking Machines Lab, positioning the company as a research-driven AI startup focused on developing advanced, high-impact AI systems. The venture immediately drew intense attention from the tech world, not only because of its founders’ pedigrees but also due to the scale of its ambitions.
That attention translated quickly into capital.
In July, just months after its formation, Thinking Machines closed a staggering $2 billion seed round—one of the largest seed financings ever recorded in Silicon Valley. The round was led by Andreessen Horowitz and included participation from Accel, Nvidia, AMD, Jane Street, and other prominent investors. The financing valued the startup at approximately $12 billion, placing it among the most valuable private AI companies almost overnight.
Such a valuation reflected investor confidence in Murati’s vision and in the ability of her team to compete at the highest levels of AI research and commercialization.
?Cracks Beneath the Surface
Despite its financial success and high-profile team, Thinking Machines has now seen the departure of multiple co-founders within a relatively short timeframe.
In addition to Zoph and Metz, the company previously lost co-founder Andrew Tulloch, who left in October to join Meta. Tulloch’s exit was initially viewed as an isolated case, but in light of the most recent developments, it now appears to be part of a broader pattern of turnover at the top.
Wired has reported that the split between Zoph and Thinking Machines Lab was not amicable—a claim that neither Murati nor Zoph has publicly confirmed. Still, observers have noted that Murati’s understated announcement, which lacked personal acknowledgment or praise for Zoph’s contributions, stands in contrast to the more celebratory tone often used when co-founders part on good terms.
TechCrunch has reached out to both Thinking Machines and OpenAI for comment, but neither organization has publicly elaborated on the reasons behind the departures as of this writing.
OpenAI’s Enduring Pull on AI Talent
The return of multiple senior researchers to OpenAI highlights the organization’s continued ability to attract—and re-attract—top-tier talent, even as former leaders launch competing ventures.
OpenAI itself has experienced a steady stream of departures over the past two years, with several co-founders and senior researchers leaving to start or join rival companies. John Schulman, for example, departed OpenAI for Anthropic in August 2024, before later joining Thinking Machines Lab as Chief Scientist at its launch in February of the following year.
This fluid movement of talent is not uncommon in Silicon Valley, particularly in fast-moving sectors like artificial intelligence. However, the simultaneous return of multiple co-founders to OpenAI—especially from a well-funded startup led by a former OpenAI executive—is unusually striking.
It suggests that OpenAI remains a uniquely compelling environment for AI researchers, whether due to its access to vast computational resources, its central role in shaping AI policy and deployment, or its continued influence over the direction of frontier AI research.
Implications for Thinking Machines Lab
The departure of two co-founders, including the CTO, less than a year after the company’s founding could be perceived as a significant setback for Thinking Machines Lab.
Startups at such an early stage often rely heavily on the cohesion and shared vision of their founding teams. Losing key figures can disrupt research roadmaps, slow execution, and raise concerns among employees, partners, and investors.
That said, Thinking Machines is far from defenseless. The appointment of Soumith Chintala as CTO provides the company with a highly respected technical leader whose influence in the AI ecosystem is substantial. Chintala’s experience and reputation could help stabilize the organization and reassure stakeholders about its technical direction.
Moreover, the company’s massive funding war chest gives it considerable flexibility to recruit new talent, invest in infrastructure, and weather periods of internal transition that would be fatal for less well-capitalized startups.
A Broader Industry Signal
Beyond the immediate implications for Thinking Machines and OpenAI, these moves reflect deeper trends shaping the AI industry.
As competition intensifies among AI labs, the battle for elite researchers is becoming increasingly fluid and high-stakes. The boundaries between companies are porous, and professional allegiances can shift rapidly in response to strategic alignment, leadership dynamics, and access to resources.
The episode also illustrates the challenges faced by even the most well-funded startups when attempting to build cohesive, long-term teams in a market where established players like OpenAI, Meta, and Google can offer unparalleled scale and influence.
An Unfinished Story
For now, both companies are moving forward.
OpenAI strengthens its research bench with the return of experienced leaders who know its culture and systems intimately. Thinking Machines Lab, under Murati’s leadership, enters a new phase—one that will test its ability to adapt, execute, and justify the extraordinary confidence investors have placed in it.
Whether these departures represent a temporary disruption or a more fundamental challenge to Thinking Machines’ long-term vision remains to be seen. What is clear, however, is that in the rapidly evolving world of artificial intelligence, even the most promising startups are not immune to sudden and dramatic shifts.
And as the lines between competition, collaboration, and return continue to blur, the movement of talent may prove just as consequential as the technology itself.
