Skip to main content
The Markets by Proactive
Go to Proactive UK
Proactive UK has moved. Proactive’s coverage of London’s small caps continues on proactiveinvestors.com Go there →
Advertisement
The Markets
by Proactive
Proactive UK has moved.
Coverage of London’s small caps continues on proactiveinvestors.com
Go to Proactive UK
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK
Advertisement
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK

Tech

AI robot startup Physical Intelligence raises $400m from Bezos, OpenAI

Physical Intelligence, a developer of 'robot brains', has been valued at $2 billion as it completed a $400 million fundraising from the likes of Jeff Bezos, OpenAI and Joshua Kushner's Thrive Capital.

The New York Times reported the funding round for the San Francisco-based startup, which saw its valuation jump from around $400 million in March this year, when it raised $70 million of seed capital.

Physical Intelligence was founded by Karol Hausman, Chelsea Finn, Sergey Levine and Brian Ichter, who left Google DeepMind to start the company with Lachy Groom, a Silicon Valley investor and an early leader of fintech Stripe.

Hausman told the NYT that the startup is not making a brain for any particular robot but "a single generalist brain that can control any robot".

In a blog post that announced the creation of its first model, π0 (pi-zero) last month, Physical Intelligence said it has developed the general-purpose model as a first step toward a long-term goal of developing "artificial physical intelligence, so that users can simply ask robots to perform any task they want, just like they can ask large language models (LLMs) and chatbot assistants".

excited to share what we've been up to soon! pic.twitter.com/TcmKXw4mqD

— Physical Intelligence (@physical_int) October 30, 2024

Unlike LLMs such as ChatGPT or Google's Gemini, the AI model can process images, text and actions and "acquires physical intelligence by training on embodied experience from robots, learning to directly output low-level motor commands via a novel architecture".

It can control other different robots, and can either be prompted to carry out a desired task, or fine-tuned to specialize in more challenging scenarios.

Advertisement
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK