Thinking Machines Lab cofounders return to OpenAI
Thinking Machines Lab cofounders return to OpenAI — and that’s reshaping the competitive map for frontier AI research. According to reporting on an internal memo, cofounders Barret Zoph and Luke Metz are leaving Thinking Machines Lab (TML) to rejoin OpenAI, with another staffer, Sam Schoenholz, also returning to the company. The moves come as TML faces rapid senior turnover, underscoring the intense competition among Big Tech and startups for top AI researchers [1].
Background: a fast‑rising lab with outsized funding
TML was founded in early 2025 by former OpenAI CTO Mira Murati and several ex–OpenAI researchers. The company quickly assembled elite talent and reportedly raised about $2 billion in seed funding, with additional multibillion‑dollar capital now being sought at a high valuation. The lab has ambitious research and product goals within a crowded frontier‑model landscape, elevating the stakes of leadership retention and execution [1][4].
Timeline of departures and key details
- Andrew Tulloch, a TML cofounder, left earlier to rejoin Meta after an 11‑year prior stint there and subsequent time at OpenAI. Reports suggest Meta offered a highly lucrative, multi‑year compensation package, though exact figures are disputed [1][2][3].
- Zoph and Metz are departing to rejoin OpenAI, according to an internal memo. Schoenholz is also returning to OpenAI, adding to the broader reshuffle of high‑end AI research talent [1].
- OpenAI applications chief Fidji Simo told staff that Zoph informed Mira Murati he was considering leaving and was then fired by TML the same day. A source close to TML alleged Zoph shared confidential information with competitors, which has not been independently verified. OpenAI stated it does not share Murati’s concerns [1].
Allegations and verification: confidentiality claims and responses
The allegation that Zoph shared confidential information remains unverified. A source close to TML made the claim, but the report notes it has not been independently confirmed. OpenAI said it does not share Murati’s concerns, highlighting the limits of public information available so far. As further details emerge, both the veracity of the allegation and any potential legal or policy ramifications will bear close watching [1].
Thinking Machines Lab cofounders return to OpenAI: why it matters
The departures signal intensifying AI talent competition between platform incumbents and startups. For TML, leadership churn at this stage could complicate execution on ambitious R&D and product roadmaps. For incumbents, bringing back experienced researchers consolidates capabilities, accelerates internal roadmaps, and potentially reduces competitive risk from fast‑moving challengers [1].
From an investor’s perspective, the headline figure—Thinking Machines Lab $2 billion seed funding—amplifies the importance of retention and governance. Large early rounds magnify pressure to hit milestones; sustained turnover in core technical leadership can increase delivery risk and affect future fundraising dynamics [1][4].
Compensation and competition: the Meta offer and the arms race
Reports describe Meta’s package to Tulloch as highly lucrative and multi‑year, with one widely cited post suggesting it could be worth up to $1.5 billion over six years, though that figure is disputed in other coverage. A TML spokesperson has described Tulloch’s exit as driven by personal reasons. Regardless of the exact number, the throughline is clear: compensation and long‑term incentives are central to mobility in the frontier‑model talent market, and Big Tech is willing to bid aggressively to secure key leaders [1][2][3].
Practical takeaways for startup leaders and investors
- Align incentives to milestones: Calibrate equity and retention awards to mission‑critical R&D and product milestones to reduce the allure of external offers during pivotal phases [1][4].
- Strengthen disclosure and IP controls: Ensure clear protocols around confidential information and staff transitions, maintaining both compliance rigor and fairness in employee relations [1].
- Scenario‑plan leadership churn: Establish succession plans and knowledge‑transfer routines so roadmaps remain resilient despite unexpected departures [1][4].
- Communicate with stakeholders: Proactive updates to teams and investors can preserve trust as leadership dynamics evolve in a competitive hiring market [1].
As the market absorbs these moves, companies seeking to build or retain advanced AI teams should keep a close eye on compensation structures and culture—factors that materially influence retention even in well‑funded environments. For additional frameworks on operational readiness, see our playbooks.
What to watch next
- Additional exits or hires across TML, OpenAI, and Meta as competition accelerates [1][3][4].
- TML’s fundraising progress and any valuation changes amid leadership turnover [1][4].
- Official statements or policy updates from the companies involved as more information is vetted publicly [1].
For broader industry context on how top AI labs are scaling research and products, readers can consult major publications’ ongoing coverage, such as The New York Times (external), alongside the original reporting cited below.
Sources
[1] Two Thinking Machines Lab Cofounders Are Leaving to Rejoin …
https://www.wired.com/story/thinking-machines-lab-cofounders-leave-for-openai/
[2] Meta’s $1.5B Offer: Thinking Machines Lab Co-Founder Leaves …
https://www.linkedin.com/pulse/metas-15b-offer-thinking-machines-lab-co-founder-leaves-aj-green-sckjc
[3] Meta Taps Thinking Machines Co-Founder to Boost AI Expertise
https://www.pymnts.com/artificial-intelligence-2/2025/meta-taps-thinking-machines-co-founder-to-boost-ai-expertise/
[4] Inside Thinking Machines Lab, Mira Murati’s New AI Startup | Built In
https://builtin.com/articles/what-is-thinking-machines-lab
