Generalist

Generalist is a robotics foundation-model company building what it calls general intelligence for the physical world. It was founded by Pete Florence and Andy Zeng, both formerly of Google DeepMind's robotics team, together with Andrew Barry, previously an engineer at Boston Dynamics.

The company's models are named GEN-1 and GEN-1.5. The pitch is one-shot learning: rather than collecting a large task-specific dataset and fine-tuning, GEN-1.5 is designed to pick up a new manipulation task in context from a single demonstration of roughly three to twelve seconds, or to adapt with one to ten gradient steps on minutes of data. GEN-1 supports a broad range of end effectors, so the same model drives different robot hands and grippers rather than being tied to one hardware platform. Generalist reports an average success rate of about 59 percent across a ten-task evaluation, which is a useful reminder that one-shot generalization in the physical world is an early capability rather than a solved one.

Funding has moved quickly. A round of about $400 million, led by Radical Ventures with participation from NVIDIA and Bezos Expeditions, valued the company at roughly $2 billion in June 2026. Two months later TechCrunch reported, citing sources, that an extension of about $200 million led by 8VC had taken the Series B to roughly $600 million and the valuation to about $3 billion. Union Square Ventures and the researcher Fei-Fei Li are also among its backers.

Access is the thing to understand before evaluating it. There are no public weights, no API, no published pricing and no self-serve product. Generalist runs its models on its own robot fleet and data engine, and works with a small number of customers through direct partnership. It is a research company with a commercial pipeline rather than a product a reader can pick up and try.