
Robotics startup Generalist reaches three billion dollar valuation
Generalist has raised nearly two hundred million dollars in fresh capital, bringing its total Series B funding to six hundred million dollars. The company is developing foundation models designed to allow robots to learn new tasks from short video demonstrations.
Published by Jin · 2 min read · 26 AUG 2026
Generalist, a robotics startup founded in 2024, has reached a valuation of three billion dollars following an additional funding extension led by 8VC. According to regulatory filings, the recent capital injection totals nearly two hundred million dollars.
This funding serves as an extension of a four hundred million dollar Series B round led by Radical Ventures, which was initially announced in June at a two billion dollar valuation. With this latest addition, the total funding for the round has reached six hundred million dollars.
Origins and Technology
The startup was established by former Google DeepMind researchers Pete Florence and Andy Zeng, alongside former Boston Dynamics engineer Andrew Barry. Early backers include Nvidia, Union Square Ventures, Bezos Expeditions, and AI researcher Fei-Fei Li.
Generalist is currently focused on developing an AI foundation model capable of operating across various robotic hardware. The company recently released its Gen 1.5 model, which aims to allow robots to master new tasks using video demonstrations ranging from three to twelve seconds in length. The team is collaborating with a select group of customers to refine the model for specific practical applications.
Market Context
The broader sector for robotic foundation models continues to attract significant venture capital. Competitors in the space include Physical Intelligence, Skild AI, and Genesis AI, reflecting strong investor interest in creating generalized software brains for hardware systems.
While some investors believe the robotics industry may be approaching a significant capability milestone similar to early language models, others caution that the lack of vast internet-scale training data for physical hardware means truly general robotics models may require further time to develop.
Source — Original announcement ↗
Worth a read?
Comments · 0