Physical Intelligence Nears $1 Billion Round, Doubling Valuation in Four Months

Physical Intelligence is trying to raise another billion dollars, this time at a valuation north of $11 billion, Bloomberg reported Friday. That would double what the two-year-old robotics startup was worth just four months ago, when it closed a $600 million round at $5.6 billion. The deal talks are early. Terms could change.

Peter Thiel joins the cap table

Peter Thiel's Founders Fund wants in, according to people familiar with the discussions. So does Lightspeed Venture Partners. Thrive Capital and Lux Capital, both existing investors, are in talks to come back for more. The cap table is already stacked. Jeff Bezos. CapitalG. Index Ventures. T. Rowe Price. OpenAI and Sequoia Capital showed up for the seed round.

Close this round and the company will have pulled in more than $2 billion since launching in March 2024. Seventy million in seed money. A $400 million Series A that pegged the valuation at $2.4 billion. Then a $600 million Series B, led by Alphabet's CapitalG with Bezos writing another check, that pushed it to $5.6 billion. Each round roughly doubled what the company was worth. This one would too.

ChatGPT, but for robots

So where is the money going? CEO Karol Hausman left Google DeepMind in 2024 and brought five colleagues from Stanford and Berkeley with him. The whole company exists to answer one question: can a single AI model learn to control any robot doing any task?

π0 (pi-zero) dropped in October 2024 as the base layer for robot control. π0.5 came after, pushing beyond arms bolted to tables and into robots that move through rooms.

Chatbots hallucinate and the worst that happens is a wrong answer on a screen. A robot hallucinates and a steel arm puts a dent in the wall. The models process camera feeds in real time. Someone shoves a chair two feet to the left and the system has to notice, instantly. A coffee mug looks nothing like a roll of tape to a human, but a vision model can choke on the difference. Physical Intelligence trains across dozens of robot types and hundreds of tasks, feeding vision and language data into proprietary architectures. The problem they're chasing has haunted manufacturing for decades. Build a machine for one job in one setting and it locks up the moment anything shifts.

Sergey Levine, one of the co-founders, put it simply when TechCrunch visited in January. "Think of it like ChatGPT, but for robots." Eighty people, give or take, work out of San Francisco. The robotic arms they train fold clothes, snap boxes together, make coffee. Three-hour test sessions, three minutes per task on average. Getting enough training data to feed the models has been one of the company's toughest problems, leaders said during the Series A fundraise.

But commercialization? Nowhere on the roadmap. Lachy Groom, another co-founder, has been blunt. "There's no limit to how much money we can really put to work," he told TechCrunch. "There's always more compute you can throw at the problem." No product. No revenue target. No timeline for either. That kind of candor would send most investors running. In physical AI, it lands an $11 billion valuation.

A crowded and expensive race

Physical Intelligence is not the only company burning through capital on this bet. The broader physical AI sector is swallowing money at a pace that makes even the LLM funding frenzy look measured.

Skild AI, building a robot-agnostic model called Skild Brain, was in talks last December for over $1 billion at a $14 billion valuation from SoftBank and Nvidia. Figure's humanoid robot hit a $39 billion valuation in September after raising more than $1 billion.

And this week, Periodic Labs entered deal talks at $7 billion, up from $1.3 billion in its seed round six months ago. Young AI labs commanding multi-billion valuations is no longer the exception. It is the market.

Jensen Huang's GTC keynote last week poured fuel on the fire. Nvidia disclosed a $1 trillion order book for its Blackwell and Vera Rubin chips through 2027. That is double what the company projected a year earlier. Physical AI, Huang told the audience, is one of five strategic pillars going forward.

The $11 billion question

What separates Physical Intelligence from this crowd? One model. Every robot. Every task. At least that's the pitch. Figure builds its own humanoid hardware. Skild ships partnerships with LG and HP Enterprise. Physical Intelligence ships research. Strong research, from people who helped build the field. But research all the same.

Eighty employees. No commercial product. No revenue. Possibly worth $11 billion by the time you read this. Investors are betting, with real conviction, that whoever cracks general-purpose robot intelligence will own something measured in trillions. Maybe they're right. But the robots still have to fold the laundry, brew the coffee, and work a kitchen counter somebody forgot to wipe down. That's the part you can't put in a pitch deck.

Shield AI Raises $2 Billion at $12.7 Billion Valuation, Plans Aechelon Acquisition

Shield AI said Thursday it raised two billion dollars in fresh capital, a figure that values the autonomous drone maker at $12.7 billion, according to the company and the New York Times. That is more

The Implicator

Yann LeCun Raises $1 Billion for AMI Labs in Europe's Largest Seed Round

Advanced Machine Intelligence Labs, the Paris-based startup founded by former Meta chief AI scientist Yann LeCun, announced Tuesday it raised $1.03 billion in seed funding at a $3.5 billion pre-money

The Implicator

Sitegeist Raises €4M to Automate Concrete Repair With Autonomous Robots

Munich startup Sitegeist has raised €4 million in a pre-seed round to deploy autonomous robots that strip deteriorated concrete from aging bridges, tunnels, and parking structures, the company announc

The Implicator

Maria Garcia

Los Angeles

Bilingual tech journalist slicing through AI noise at implicator.ai. Decodes digital culture with a ruthless Gen Z lens—fast, sharp, relentlessly curious. Bridges Silicon Valley's marble boardrooms, hunting who tech really serves.