Back to News
Technology
Aug 26, 20260 views2 min read

Generalist Robotics Raises 200 Million More, Reaches 3 Billion Valuation

Robotics AI startup Generalist closed a 200 million dollar funding extension in August 2026, led by 8VC, bringing its total Series B to 600 million dollars and its valuation to 3 billion dollars. The company, founded by former Google DeepMind and Boston Dynamics researchers, builds AI foundation models that let robots learn new tasks from short video demonstrations. Its latest model, Gen-1.5, achieved an 83 percent task completion rate with a few examples.

Generalist Robotics Raises 200 Million More, Reaches 3 Billion Valuation
Source:TechCrunch

Robotics AI startup Generalist closed a $200 million funding extension in late August 2026, led by 8VC, bringing its total Series B capital to $600 million and its valuation to $3 billion.

The extension builds on a $400 million Series B round announced in June 2026, led by Radical Ventures. The company's valuation rose from $2 billion at the time of the initial round to $3 billion following the August close.

Generalist was founded in 2024 by Pete Florence and Andy Zeng, former Google DeepMind researchers, and Andrew Barry, a former Boston Dynamics engineer. The startup focuses on building AI foundation models, often described as "robot brains," rather than physical hardware.

The company's latest model, Gen-1.5, released in August 2026, enables robotic arms to master new tasks based on video demonstrations lasting between 3 and 12 seconds. Users can teach robots by demonstrating tasks with their own hands or using simulated robot clips. In testing, the model achieved a 59 percent task completion rate with a single example, rising to 83 percent with a few additional examples.

Key investors include Nvidia, Bezos Expeditions, Union Square Ventures, Spark Capital, and AI researcher Fei-Fei Li.

Generalist operates in a competitive field alongside Skild AI, valued at $14 billion, and Physical Intelligence, valued at $11 billion. The surge in investment reflects a broader industry bet on the potential for robots to perform diverse tasks without requiring explicit programming for every specific action.

Industry analysts note that fully general-purpose robotics models may still be years away, as robots cannot be trained on the vast scale of internet data available to large language models.