Generalist AI just closed roughly $200 million in fresh funding, led by 8VC, pushing the San Mateo startup’s total capital raised past the $650 million mark. The round values the company north of $2 billion, a figure it first hit just months ago during its $400 million Series B in June 2026.
For a company founded in 2024, that’s a staggering trajectory. Then again, when your founding team includes alumni from both DeepMind and Boston Dynamics, venture capitalists tend to move quickly.
Building brains for robots
Generalist AI is working on what the industry calls “embodied AI foundation models.” Think of it as giving robots the kind of flexible, general-purpose intelligence that large language models gave chatbots, except instead of generating text, these systems control physical machines performing real-world tasks.
The company has been shipping at a rapid clip. It released GEN-0 in November 2025, followed by GEN-1 in April 2026. GEN-1.5 is currently in development.
The models are trained on hundreds of thousands of hours of proprietary physical interaction data. According to the company, the results are striking: 99% reliability on diverse dexterous tasks and execution speeds up to 3x faster than previous benchmarks.
The team behind the models
Generalist AI was co-founded by CEO Pete Florence, Chief Scientist Andy Zeng, and CTO Andrew Barry. Florence and Zeng both came from DeepMind. Barry’s experience at Boston Dynamics adds deep expertise in the mechanical and control systems side of the equation.
The June 2026 Series B was led by Radical Ventures. That $400 million round valued Generalist AI at $2 billion post-money, and the new $200 million infusion from 8VC suggests the valuation has only climbed from there.
A crowded but well-funded race
The most dramatic comparison point is Physical Intelligence, a competitor that has reportedly achieved a valuation exceeding $11 billion.
Manufacturing, logistics, and supply chain operations are the obvious first targets, but the long-term ambitions extend further. Healthcare, agriculture, construction, and household assistance all represent massive addressable markets where adaptable robots could eventually operate.