Meta recently lost AI researcher Andrew Tulloch to Anthropic after less than a year back at the company. Tulloch will join the Claude-maker’s inference and performance team next week. The move follows Meta’s intense effort to recruit him during its superintelligence push. Reports put his earlier Meta package at up to $1.5 billion over six years.
Tulloch told colleagues about his departure shortly after Meta launched Muse this week. The AI researcher reportedly waited for the assistant’s launch before leaving Meta. Anthropic confirmed his new role, where he will focus on AI model performance and efficiency. The move gives Anthropic another highly experienced researcher as leading labs compete for scarce talent.
Tulloch spent 11 years at Meta before moving to OpenAI in 2023. He later helped train major models including GPT-4o, GPT-4.5 and o3. Tulloch then co-founded Thinking Machines Lab with former OpenAI executive Mira Murati. Meta eventually brought him back in October 2025.
At Meta, Tulloch worked inside TBD, a team supporting Mark Zuckerberg’s superintelligence strategy. Alexandr Wang leads Meta’s broader Superintelligence Labs organization. Tulloch’s work focused on large-scale machine learning systems and advanced AI development.
The reported $1.5 billion figure needs important context. Sources described a potential six-year package linked to bonuses and Meta stock performance. Meta spokesperson Andy Stone rejected the figure as ‘inaccurate and ridiculous.’ Meta never disclosed the final compensation terms Tulloch accepted.
His move also shifts attention toward AI infrastructure rather than headline model releases. Anthropic’s inference and performance team works on improving how efficiently models operate. Better efficiency can reduce computing costs while improving speed and large-scale product performance.
The departure also exposes the limits of massive compensation packages in retaining AI specialists. Research autonomy, technical influence, computing access and team direction can shape senior researchers’ decisions. Meta, Anthropic, OpenAI and Google continue competing for the same small pool of elite AI talent.
Tulloch’s exit therefore carries a wider message for the AI industry. Even extraordinary offers cannot guarantee long-term loyalty when researchers have several powerful career options. The escalating AI talent war now extends beyond salaries into technology, research culture and strategic influence.
Also Read: Meta Launches Muse AI Agent, Can Send Emails, Book Travel: How It Works