Mira Murati Fires Barret Zoph Thinking Machines Lab Drama Sends Shockwaves Through AI Industry Unleashed

Mira Murati fires Barret Zoph Thinking Machines Lab in a messy split that exposed fractures at one of 2025’s hottest AI startups. The former OpenAI CTO turned CEO axed her co-founder and CTO over performance, conduct, and an alleged undisclosed workplace relationship. Hours later, Zoph landed back at OpenAI alongside other defectors.

• What happened: Murati announced the parting on X, naming Soumith Chintala as new CTO. Internal messaging cited unethical conduct and performance gaps.
• Why it matters: Thinking Machines Lab raised a massive $2 billion seed at $12 billion valuation. Losing key talent early signals leadership and culture risks in fast-moving AI ventures.
• The fallout: Zoph disputed the firing, claiming it followed his notice of potential departure. OpenAI welcomed him back without sharing Murati’s concerns.
• Broader picture: Multiple co-founders have left, raising questions about retention at post-OpenAI labs.
• Lesson for builders: High-stakes AI teams live and die by trust, clear policies, and execution speed.

This clash isn’t just gossip. It highlights how personal dynamics and power struggles can torpedo even well-funded AI efforts. Here’s the full breakdown.

Background on Mira Murati, Barret Zoph, and Thinking Machines Lab

Mira Murati left OpenAI in late 2024 to launch Thinking Machines Lab in early 2025 with several ex-colleagues, including Barret Zoph as co-founder and CTO. The startup quickly scored eye-watering funding from Andreessen Horowitz, Accel, Nvidia, and others. Their goal? Push frontiers in customizable AI models and tooling.

Zoph brought serious cred from OpenAI’s research side, particularly vision and architecture work. Murati steered the ship as CEO. For months, things looked golden. Then cracks appeared.

Tensions reportedly built around an office relationship Zoph had with a colleague. Sources say Murati confronted him about it last summer. What started as a conversation snowballed into productivity concerns, role changes, and eventually termination.

The kicker is how fast everything unraveled. One day Zoph is CTO at a $12B startup. Days later, he’s back at OpenAI.

Timeline of Mira Murati Fires Barret Zoph Thinking Machines Events

Early tensions simmered through 2025. Murati moved Zoph toward an individual contributor role after the relationship came to light. Performance gripes mounted.

In January 2026, things boiled over. Zoph allegedly told Murati he was eyeing other opportunities. A contentious meeting followed. Murati terminated his employment, citing conduct and performance. She posted publicly: “We have parted ways with Barret Zoph. Soumith Chintala will be the new CTO.”

OpenAI announced his return within the hour, along with Luke Metz and Sam Schoenholz. Other departures, like Andrew Tulloch to Meta earlier, added to the exodus narrative.

Event	Date (approx.)	Key Details	Impact
Startup Founded	Early 2025	Murati, Zoph, Metz + team leave OpenAI	High talent concentration
Major Funding	July 2025	$2B seed at $12B valuation	Massive resources secured
Relationship Confrontation	Summer 2025	Alleged undisclosed office romance	Trust erosion begins
Role Adjustment	Late 2025	Zoph shifted to IC role	Productivity questions rise
Firing Announced	January 2026	Murati cites conduct/performance	Public drama, talent flight
OpenAI Re-hire	Same day	Zoph + others return	Signals ongoing OpenAI pull

This table cuts through the noise. Leadership transitions in AI rarely stay quiet.

What Led to the Decision: Performance, Conduct, and Culture

In my experience building and advising tech teams, firing a co-founder is nuclear. You only pull that trigger when trust is gone and execution suffers.

Murati’s internal note highlighted performance shortfalls and conduct issues. The undisclosed relationship allegedly violated norms, especially since Zoph had pushed for the hire. He initially denied it, per reports.

Here’s the thing: In hyper-competitive AI, where IP moves fast and teams are small, personal entanglements can poison decision-making. Did it affect promotions, priorities, or information flow? Those questions matter.

Zoph pushed back, framing it as retaliation after he signaled departure. OpenAI’s quick welcome suggests they saw no red flags on their end.

Rhetorical question: When does a personal issue become a business risk worth firing over?

Lessons for AI Startups: Navigating Leadership Drama

Mira Murati fires Barret Zoph Thinking Machines case offers a masterclass in what happens when founder alignment breaks.

Seasoned operators know this: Document everything. Have clear HR policies on relationships from day one. Conduct regular performance reviews—even with co-founders. When tensions rise, address them early before they fester into public spectacles.

If I were advising a similar lab today, I’d implement quarterly anonymous feedback and third-party mediation options for founder disputes. Prevention beats damage control.

Step-by-Step Action Plan for Beginners in AI Leadership

New to managing high-talent tech teams? Follow this:

1. Set expectations upfront — Draft a simple founder agreement covering roles, exits, and conflict resolution.
2. Build transparent culture — Regular all-hands, written updates, and open channels reduce rumors.
3. Handle relationships proactively — Create clear policies on workplace romances and disclosures. Enforce consistently.
4. Track performance objectively — Use measurable KPIs. Document gaps with specific examples.
5. Plan for departures — Have succession ready. Murati moved fast with Soumith Chintala—smart move.
6. Communicate externally with care — Brief, factual statements protect the brand without fueling drama.

Start small. Review your current team charter this week.

Common Mistakes & How to Fix Them

Mistake 1: Ignoring early red flags. Founders often delay tough talks. Fix: Schedule structured check-ins monthly. Address issues in writing.

Mistake 2: Blurring personal and professional lines without guardrails. The alleged relationship fallout shows the risk. Fix: Adopt standard tech industry policies early. Train managers on them.

Mistake 3: Poor succession planning. Losing a CTO can stall momentum. Fix: Cross-train key people and maintain a bench of talent.

Mistake 4: Letting drama leak publicly before internal alignment. Fix: Control the narrative with prepared statements and unified messaging.

Mistake 5: Over-relying on star talent without systems. AI moves too fast for hero culture. Fix: Build processes that survive individual exits.

Key Takeaways

• Mira Murati fires Barret Zoph Thinking Machines exposed how quickly founder dynamics can shift in AI.
• Trust and performance trump past glory when stakes are this high.
• Workplace relationship policies are non-negotiable in tight-knit teams.
• Fast funding doesn’t guarantee smooth sailing—execution and culture do.
• OpenAI’s talent gravity remains strong despite spin-offs.
• Clear communication and documentation save headaches during exits.
• Succession moves, like naming Soumith Chintala, signal resilience.
• Every drama teaches: prioritize alignment over speed in hiring.

Mira Murati fires Barret Zoph Thinking Machines reminds us the AI race isn’t just about models—it’s about people who build them. Nail leadership and retention, and you stay in the game. Slip, and even $12 billion won’t save you.

Next step: Audit your team’s current policies on conduct and performance. Do it before the next crisis hits.

FAQs