Andrew Tulloch, one of Silicon Valley’s most sought-after AI engineers, has left Thinking Machines Lab, the $12 billion startup he co-founded with former OpenAI CTO Mira Murati, to join Meta. The move follows months of speculation surrounding Tulloch’s next step – and comes after he reportedly declined a Meta offer earlier this year worth up to $1.5 billion over six years.
Tulloch’s departure marks a significant moment in the escalating talent wars among major AI players. According to sources familiar with the matter, his decision to leave Thinking Machines Lab was made for “personal reasons,” though neither company disclosed the terms of his new role. Meta confirmed the hire but declined to comment on compensation or responsibilities.
Tulloch began his career at Goldman Sachs before transitioning into the tech sector. In 2012, he joined Facebook (now Meta), where he spent more than a decade building machine learning and large-scale infrastructure systems. His work contributed to early versions of Meta’s AI architecture, forming the foundation for the company’s later advances in generative and recommendation AI.
In 2023, Tulloch joined OpenAI, where he helped design internal systems that scaled the training of large models, including GPT-4. These systems optimized OpenAI’s hardware utilization, cutting training times and operational costs, and became a cornerstone of the company’s expansion into large-scale inference workloads.
After leaving OpenAI, Tulloch co-founded Thinking Machines Lab in early 2025 with Murati. The startup quickly became one of the most closely watched newcomers in the AI space, raising $2 billion in funding and achieving a $12 billion valuation within months. Its flagship product, Tinker, launched as an API platform for fine-tuning and deploying large language models tailored to enterprise needs.
The company’s rapid rise drew interest from major tech firms. Mark Zuckerberg reportedly attempted to acquire Thinking Machines Lab earlier this year but was rebuffed. When the acquisition failed, Meta pursued Tulloch directly, offering him a compensation package estimated at $1.5 billion, tied to performance and stock options. Tulloch turned it down, and at the time, no senior staff left the startup.
His eventual move to Meta, announced this week, surprised many industry insiders. People close to the company described the decision as a “quiet personal shift” rather than a strategic exit. A spokesperson for Thinking Machines Lab said Tulloch “chose to step away to pursue a different path,” while reaffirming that Murati remains CEO and that the company’s roadmap remains unchanged.
Meta has not yet revealed Tulloch’s official title or responsibilities, but insiders suggest he will join Meta’s AI Research and Infrastructure group, focusing on large-model training efficiency and scaling.
Tulloch’s move underscores the intensity of competition among Big Tech firms vying for top AI talent – where billion-dollar offers, equity-heavy packages, and strategic hires are becoming standard in the race to dominate artificial intelligence’s next frontier.
