Mira Murati helped lead the team behind ChatGPT, DALL-E, and Codex as OpenAI’s CTO. In September 2024, she walked away. Five months later, she launched Thinking Machines Lab and raised $2 billion at a $12 billion valuation in a seed round led by Andreessen Horowitz, before shipping a single product.
Investors were not betting on technology. They were betting on who left the building.
From Product Operator to AI Power Broker
The distinction matters. Murati did not single-handedly build ChatGPT, as some profiles have claimed. She joined OpenAI in June 2018 as VP of Applied AI and Partnerships, was promoted to CTO in May 2022, and oversaw the research, product, and safety teams during the period when OpenAI shipped its most consequential products.
Her background set her apart from most AI leaders. Before OpenAI, she worked as a senior product manager at Tesla on the Model X program and as VP of Product and Engineering at Leap Motion. The common thread was translating complex technology into products people actually use.
When OpenAI’s board fired Sam Altman in November 2023, Murati was named interim CEO for approximately two days before the situation resolved with Altman’s reinstatement. The episode revealed how central she had become inside the organization, and according to reports in Fortune, it also complicated her relationship with Altman. Ten months later, she announced her departure.
The $2 Billion Bet on a Person
Thinking Machines Lab’s seed round tells a specific story about what investors value in AI right now. Andreessen Horowitz led the $2 billion round at a $12 billion valuation, with NVIDIA, AMD, Accel, and Jane Street Capital among the participants. Crunchbase data confirms it as the largest seed round in history.
What the investors bought was not a product or a prototype. They bought Murati’s track record and the team she assembled, including John Schulman (an OpenAI co-founder who joined as Chief Scientist), Barret Zoph (OpenAI’s VP of Research who became CTO), and Lilian Weng (an OpenAI VP who joined as co-founder).
The company was structured as a benefit corporation with a governance model that gives Murati decision-making authority. In March 2026, NVIDIA committed at least 1 gigawatt of Vera Rubin compute to Thinking Machines Lab, with deployment beginning in 2027. Another billion-dollar bet placed before a product launch.
The Talent War Gets Personal
The competition for Murati’s team became its own story. According to TechCrunch and the Wall Street Journal, Meta attempted to acquire Thinking Machines Lab outright, and when that failed, offered individual engineers compensation packages reportedly worth hundreds of millions of dollars. According to Wired, one person was offered more than $1 billion. Murati told the same outlet that “not a single person has taken the offer.”
The reality proved more complicated. Andrew Tulloch, a co-founder, left for Meta in October 2025. In January 2026, CTO Barret Zoph and co-founder Luke Metz returned to OpenAI. Soumith Chintala, co-creator of PyTorch who spent thirteen years at Meta, joined as the new CTO.
These departures do not necessarily signal failure. They signal how intensely the AI industry competes for a small number of people who can build frontier AI systems.
From Vlorë to the Fortune Cover
In November 2025, Bloomberg reported that Thinking Machines Lab was in discussions to raise additional funding at a valuation of approximately $50 billion. The round never closed, but the fact that investors were willing to consider a fourfold increase in valuation within months, before a product launch, says something about what the market is actually pricing.
Meanwhile, Albania amended its national budget to invest approximately $10 million in Murati’s company. A country of fewer than three million people rewrote its fiscal plan to back a startup founded by a daughter of Vlorë, the port city where Murati grew up during the civil unrest of the late 1990s.
TIME named Murati one of the 100 Most Influential in AI in 2024. CNBC named her a 2026 Changemaker. Fortune put her on their Most Powerful Women list. For a girl who left Albania on a scholarship at sixteen, the trajectory is extraordinary. For the AI industry, it is a case study in what happens when the people who built the most important products decide to build something else.
The Pattern Behind One Story
Murati’s story is not unique. Dario and Daniela Amodei left OpenAI in 2021 and built Anthropic, now valued at $61.5 billion. Ilya Sutskever, OpenAI’s co-founder and chief scientist, left in May 2024 and raised $32 billion for Safe Superintelligence before demonstrating a product. Combined, OpenAI alumni have founded companies worth over $100 billion.
This is not normal executive turnover. It is a structural shift in where AI’s center of gravity sits, with the most consequential people in the industry building competing power centers, each backed by billions committed on reputation alone.
What This Means for AI Strategy
The conventional approach to evaluating AI vendors focuses on capability benchmarks, pricing, and integration requirements. Those factors matter, but they miss something that may matter more. The AI infrastructure investment wave has poured hundreds of billions into hardware and data centers, yet the most expensive and scarce resource in AI is not silicon or electricity. It is the people who know how to build these systems.
When a key leader leaves an AI company, research priorities shift, product roadmaps change direction, and institutional knowledge walks out the door. The model your organization evaluated six months ago may be heading somewhere different than the roadmap suggested, not because of a strategic pivot, but because the person who defined the direction is now building something else.
For any CEO evaluating AI vendors, this pattern carries a practical question that goes beyond features and pricing: who built the model your team relies on, and are they still there?
In AI, the roadmap is partly a function of who is still in the room. OpenAI alumni have founded companies worth over $100 billion, and that should change how you evaluate the ones who remain.