Embodied AI Glossary中文Generalist AI
CommonA US embodied-AI company that trains its GEN-series foundation models on huge amounts of real-robot interaction data.
Generalist AI is a US robotics AI company founded in 2024, based in the San Francisco Bay Area (San Mateo, California), with a second team in Boston. CEO Pete Florence previously worked on PaLM-E and RT-2 at Google DeepMind, and CTO Andrew Barry came from Boston Dynamics. Its approach is to collect large amounts of real-world physical-interaction data itself and use it to train embodied foundation models. GEN-0, released in November 2025, was trained on more than 270,000 hours of real-robot data and was the company's first reported scaling-law result on a robot. GEN-1 (April 2026) focused on performing simple tasks skillfully and quickly. GEN-1.5 (August 2026) can learn a new task after watching a demonstration only about a dozen seconds long. According to its LinkedIn page, its most recent funding round raised $400 million, bringing its total funding to more than $500 million.
ExampleFine-tuned on only about an hour of robot data, GEN-1 reportedly reached a 99% success rate on tasks like folding a paper box.
- Also called
- Generalist
- Related
- GEN-0 · GEN-1 · Embodied Foundation Model · Scaling Law · Ossification · Real-Robot Data
- Sources
- Generalist AI Blog
Generalist AI - About
Generalist | LinkedIn - As of
- 2026-09