Physical Intelligence, a Specialist in Robot A.I., Raises $400 Million
The start-up raised $400 million in a funding round with investments from the likes of Jeff Bezos, Thrive Capital and OpenAI.
by https://www.nytimes.com/by/michael-j-de-la-merced · NY TimesPhysical Intelligence, an artificial intelligence start-up seeking to create brains for a wide variety of robots, plans to announce on Monday that it had raised $400 million in financing from major investors.
The round was led by Jeff Bezos, Amazon’s executive chairman, and the venture capital firms Thrive Capital and Lux Capital. Other investors include OpenAI, Redpoint Ventures and Bond.
The fund-raising valued the company at about $2 billion, not including the new investments. That’s significantly more than the $70 million that the start-up, which was founded this year, had raised in seed financing.
The company wants to make foundational software that would work for any robot, instead of the traditional approach of creating software for specific machines and specific tasks.
“What we’re doing is not just a brain for any particular robot,” said Karol Hausman, the company’s co-founder and chief executive. “It’s a single generalist brain that can control any robot.”
It’s a tricky task: Building such a model requires a huge amount of data on how to operate in the real world. Those information sets largely do not exist, compelling the company to compile its own. Its work has been aided by big leaps in A.I. models that can interpret visual data.
Among the company’s co-founders are Mr. Hausman, a former robotics scientist at Google; Sergey Levine, a professor at the University of California, Berkeley; and Lachy Groom, an investor and former executive at the payments giant Stripe.
In a paper published last week, Physical Intelligence showed how its software — called π0, or pi-zero — enabled robots to fold laundry, clear a table, flatten a box and more.
“It’s a true generalist,” Mr. Hausman said.
Physical Intelligence executives said that its software was closer to GPT-1, the first model published for chatbots by OpenAI, than to the more advanced brains that power ChatGPT.
Mr. Groom said that it was hard to predict the rate of progress: A ChatGPT-style breakthrough “could be far sooner than we expect, or it could definitely be far out.”
The field of robotics A.I. is getting crowded, with players including Skild, which is also working on general-purpose robot A.I.; Figure AI, whose backers include OpenAI and Mr. Bezos; and Covariant, which focuses on industrial applications.
Amazon has a vested interest in the industry, and has been adding more robots in its operations as it seeks to drive down costs and get orders to customers faster. Tesla also has major A.I. ambitions, with Elon Musk recently saying that the company’s humanoid robot would be “the biggest product ever of any kind.”
Explore Our Coverage of Artificial Intelligence
News and Analysis
- A national security memorandum signed by President Biden detailed how the Pentagon, the intelligence agencies and other national security institutions should use and protect A.I. technology.
- Intel was for decades Silicon Valley’s dominant chip company. But missed opportunities and poor execution left it on the sidelines in tech’s latest gold rush.
- A former OpenAI researcher who helped gather and organize the enormous amounts of internet data used to train the startup’s ChatGPT chatbot says the company broke copyright law.
The Age of A.I.
- A radio station in Poland fired its on-air talent and brought in A.I.-generated presenters. An outcry over a purported chat with a Nobel laureate quickly ended that experiment.
- Meta unveiled a set of A.I. tools for automatically generating videos, instantly editing them and synchronizing them with A.I.-generated sound effects, ambient noise and background music.
- Despite — or, perhaps, because of — the rise in artificially made images, photography is suddenly in the spotlight, in galleries in New York and beyond.