Thinking Machines Lab Implosion: Co-Founders Fired and Fleeing Back to OpenAI
Mira Murati's $12B AI startup loses three co-founders in 24 hours amid allegations of unethical conduct. OpenAI immediately rehires them, and more employees are reportedly in talks to follow.
The Story
Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati with a staggering $12 billion valuation, is imploding in spectacular fashion. Within 24 hours, the company lost three co-founders who immediately returned to OpenAI, with reports emerging that more employees are negotiating their own exits.
The drama began Wednesday when Murati announced on X that the company had "parted ways" with co-founder and CTO Barret Zoph. Just 58 minutes later, OpenAI's CEO of applications Fidji Simo announced that Zoph, along with co-founder Luke Metz and researcher Sam Schoenholz, would be rejoining OpenAI.
Technology reporter Kylie Robison first disclosed that Zoph was fired for what she described as "unethical conduct." Sources close to Thinking Machines alleged that Zoph had shared confidential company information with rival firms. According to Wired, Zoph informed Murati on Monday that he was considering leaving. By Wednesday, he was terminated.
OpenAI's response to the allegations was telling. In a memo to staff, Simo wrote that OpenAI "does not share the same concerns about Zoph as Murati," essentially validating Zoph's decision to return while dismissing the ethical concerns raised by his former employer.
The Broader Exodus
The situation deteriorated further on Thursday. Alex Heath reported that multiple Thinking Machines employees are now in discussions to follow the three co-founders back to OpenAI. Sources indicate that at least a few employees resigned immediately after a tense all-hands meeting addressing Zoph's departure.
This isn't the first major departure from Thinking Machines. Co-founder Andrew Tulloch left to join Meta in October 2024, reportedly after Mark Zuckerberg made one of his famous "Godfather offers" that couldn't be refused.
The pattern is now undeniable: Thinking Machines has lost four of its original co-founders within a year of its founding.
The Money Problem
The timing couldn't be worse for Murati. Thinking Machines has reportedly been in talks to raise more than $4 billion at a valuation between $50 billion and $60 billion. Those conversations are now likely dead in the water.
The company secured a $2 billion seed round in July 2024 led by Andreessen Horowitz, with participation from Accel, Nvidia, AMD, and Jane Street. That massive capital raise valued the pre-product company at $12 billion, a testament to the industry's willingness to bet on former OpenAI talent.
But capital alone cannot retain talent. The AI industry's talent market remains intensely competitive, and the gravitational pull of established labs like OpenAI, Anthropic, and Google DeepMind continues to destabilize newer ventures.
Structural Instability in AI Startups
This implosion reveals a fundamental fragility in the AI startup ecosystem. When your primary asset is talent, and that talent can be poached by deep-pocketed competitors offering premium compensation packages and established research infrastructure, your $12 billion valuation becomes precarious.
The Thinking Machines case also highlights the revolving door between AI labs. John Schulman, who left OpenAI for Anthropic in August 2024, later joined Thinking Machines as Chief Scientist. Now his former colleagues are returning to OpenAI, creating a bizarre circular migration pattern.
Meanwhile, the same day Thinking Machines lost three co-founders, OpenAI lost a senior safety research lead. Andrea Vallone, who specialized in how AI models respond to mental health issues, departed for Anthropic to work under alignment researcher Jan Leike, who himself left OpenAI in 2024 over concerns the company wasn't taking safety seriously enough.
What This Means
The Thinking Machines saga illustrates several uncomfortable truths about the current AI landscape:
Valuations are disconnected from fundamentals. A $12 billion valuation for a company that hadn't shipped a product, now facing a talent exodus, exposes the speculative nature of AI startup investing.
Former employer relationships create instability. When founders leave established labs to start competitors, they maintain networks that can unravel their new ventures overnight.
Capital cannot solve culture problems. Whatever happened internally at Thinking Machines, whether Zoph's departure was justified or not, the speed of the exodus suggests deeper organizational issues.
The talent war is intensifying. Every major AI lab is now actively recruiting from competitors, creating an environment where employee tenure is measured in months rather than years.
For enterprise AI buyers, this instability should raise red flags. Building critical business processes on platforms from well-funded but unstable startups carries significant vendor risk. The consolidation wave we've been predicting appears to be accelerating, with talent concentrated back into a handful of established labs.
The AI industry's "move fast and break things" ethos now extends to the companies themselves. Thinking Machines may survive this crisis, but its position as a credible OpenAI competitor has been severely damaged. For investors who valued the company at $12 billion based largely on its talent roster, the math no longer works.
This is a developing story. We will update as more information becomes available.