- Generalist AI has raised $400 million led by Radical Ventures, bringing total funding to more than $500 million.
- The company’s GEN-1 model, launched in April 2026, demonstrates 99% reliability across diverse dexterous tasks, execution up to 3x faster than prior state-of-the-art, and emergent improvisational intelligence.
- Backers include NVIDIA’s NVentures, Bezos Expeditions, Fei-Fei Li, Naval Ravikant, and Bin Lin — angels whose combined credibility spans hardware, capital, and frontier AI research.
Most robotics companies build a robot. Generalist AI is building the intelligence that goes inside all of them. That distinction is the entire bet: that the bottleneck in the coming wave of automation is not hardware, but a foundation model capable of understanding and acting in the physical world regardless of what body it inhabits.
San Francisco-based Generalist AI has raised $400 million, bringing its total raised to more than $500 million. The round was led by Radical Ventures, with new investors 8VC, Union Square Ventures, Hanabi Capital, and Norwest. All major existing investors participated significantly, including NVIDIA’s NVentures, Boldstart Ventures, Spark Capital, Bezos Expeditions, and NFDG. New angel investors named in the blog post include Bin Lin (co-founder of Xiaomi), Fei-Fei Li (founder of World Labs and the AI researcher who created ImageNet), and Naval Ravikant.
Generalist AI was founded in 2024 by Pete Florence (CEO), Andy Zeng (Chief Scientist), and Andrew Barry (CTO). Florence was a Senior Research Scientist at Google DeepMind, where he led the development of PaLM-E and RT-2 — two of the most cited embodied AI models ever published. His Google Scholar papers have been cited more than 19,000 times. Zeng co-authored PaLM-E alongside Florence at DeepMind and previously worked on scaling ChatGPT at OpenAI. Barry spent five years as a Senior Roboticist at Boston Dynamics — building Atlas, Spot, and Stretch — before joining the Broad Institute of MIT and Harvard as a machine learning scientist.
According to Boldstart Ventures, the founding team developed a proprietary data collection method using instrumented gloves that capture human manipulation data at scale — a training dataset the firm describes as “difficult for competitors to replicate.”
GEN-0 and GEN-1: from scaling laws to commercial viability
In November 2025, Generalist launched GEN-0 — the first robotics model to demonstrate scaling laws in robotics: proof that more physical experience and larger models predictably produce more capable systems. The company reached a $440 million valuation around that time, per The Deep View. In April 2026, GEN-1 showed where that path leads. According to the company’s blog post, GEN-1 achieves 99% reliability across diverse dexterous tasks, executes up to 3x faster than prior state-of-the-art, learns complex new physical skills from limited data, and demonstrates emergent improvisational intelligence — the ability to solve physical problems it was never explicitly trained on. According to The Deep View, prior state-of-the-art models achieved around 64% reliability on comparable tasks.
The model is hardware-agnostic. Generalist’s intelligence works across robotic arms, mobile robots, humanoids, and autonomous systems — a platform approach rather than optimising for a single machine or deployment context.
“Scaling robot learning creates better models, better models can do more useful physical work, and data from real businesses drives the next generation of more capable models. This is how general intelligence will emerge in the physical world: through systems that learn by acting, improve through experience, and become useful by working alongside people.” — Generalist AI blog post.
The competitive landscape
Generalist’s closest conceptual peer is Physical Intelligence (π), the San Francisco-based foundation model lab for robots. Physical Intelligence is reportedly in talks to raise $1 billion at an $11 billion valuation, led by a founding team from Google DeepMind, UC Berkeley, and Stanford. Where Physical Intelligence uses a diffusion-based architecture, Generalist emphasises real-world data at unprecedented scale and a unified model that transfers across hardware form factors. Figure AI, valued at $39 billion, takes the opposite approach — vertically integrating its own humanoid hardware and proprietary intelligence. Genesis AI raised $105 million for a robotics simulation and foundation model platform. The field is splitting between hardware-first companies and intelligence-first companies — and the unanswered question is which layer captures more value when physical AI scales.
Market context
According to Grand View Research, the global AI in robotics market is projected to grow at a CAGR of over 30% through 2033, with service robots growing at over 70% annually in 2025. Goldman Sachs projects the humanoid robot market alone will reach $38 billion by 2035 — with the intelligence layer representing the highest-margin component of that stack.
The defining question for Generalist is not whether foundation models will power the next generation of robots. GEN-1 makes a compelling case they already do. The question is whether a company building the brain for every robot can establish itself as the default before the hardware giants — Figure, Tesla, Boston Dynamics — decide that owning the intelligence layer is too strategic to outsource.
