Andrew Tulloch is leaving Meta for Anthropic after less than a year in his second stint at Meta, a high-profile move that shows how fiercely technology companies are competing for senior AI researchers. Meta reportedly offered him a compensation package worth more than $1 billion.
The move was reported on September 10, 2026. Tulloch is expected to work on inference and model performance at Anthropic. His exact transition date is not part of the announcement, so the important point is the employer change—not a precise start day.
What is known about the compensation package
The reported offer was worth more than $1 billion, but that figure describes an offer rather than money confirmed as paid. It also does not tell us how the package was structured, how much was guaranteed, or what portion, if any, Tulloch ultimately received.
That distinction matters in technology compensation. Large recruitment packages can combine salary, stock, bonuses and conditions spread over several years. A headline figure is not automatically a cash payment, annual salary or realized compensation.
From Meta to OpenAI to Thinking Machines Lab and back
Tulloch spent 11 years at Meta before leaving for OpenAI in 2023. He later co-founded Thinking Machines Lab with Mira Murati in 2025, then returned to Meta during the company’s push to expand its AI workforce.
Meta’s recruitment campaign included more than 50 AI researchers, engineers and other employees. Tulloch’s return was therefore part of a broader effort to gather scarce technical expertise—not an isolated hire made in a quiet corner of the company.
| Date or period | Career event | Organization or role |
| Earlier career | Spent 11 years at Meta before leaving in 2023. | Meta |
| 2023 onward | Worked at OpenAI after his earlier Meta tenure. | OpenAI |
| 2025 | Co-founded Thinking Machines Lab with Mira Murati. | Thinking Machines Lab |
| October 11, 2025 | His move from Thinking Machines Lab to Meta was reported. | Meta |
| September 10, 2026 | His departure from Meta for Anthropic was reported. | Anthropic |
What the move says about the AI talent war
AI companies are competing for a relatively small group of researchers and engineers who can influence model training, inference and performance. Recruiting one of those people can mean more than filling a job opening: it can affect a team’s expertise, leadership and ability to execute ambitious projects.
Meta’s reported offer above $1 billion puts a dramatic number on that competition. But Tulloch’s departure also shows the limit of using compensation headlines as a simple retention story. A major offer can attract a researcher; it cannot, by itself, explain how long that person will stay or why they later move on.
The case concerns one prominent departure. It does not show that Meta is carrying out broad layoffs of AI personnel, nor does it establish a wider exodus from the company.
No confirmed reason for the departure
Why did Tulloch leave Meta? No public reason has been given. The timing does not establish that compensation, Meta’s AI strategy, a specific product or any other single factor caused the move.
That leaves the central news clear but the motive unresolved: Tulloch is leaving Meta for Anthropic, while the public record does not explain his decision. The reported destination and expected focus on inference and performance suggest where his expertise may be applied next, but they do not reveal why he chose to move.
The bottom line
Andrew Tulloch’s move is significant because Meta recruited him back from Thinking Machines Lab during an unusually aggressive AI hiring push, then lost him again in less than a year. The reported offer above $1 billion makes the episode eye-catching; the more durable lesson is that elite AI talent remains mobile, even when companies attach extraordinary numbers to recruitment.
