The Fastest-Growing Company in Robotics AI Just Doubled Its Valuation

Physical Intelligence, the San Francisco-based robotics AI startup founded by former Google DeepMind researchers, is in talks to raise approximately $1 billion in new funding at a valuation exceeding $11 billion, according to Bloomberg. If the deal closes, it would nearly double the company's $5.6 billion valuation from just four months ago — and push its cumulative funding past the $2 billion mark.

In a market where AI robotics has become the hottest investment category of 2026, Physical Intelligence stands out not because it builds robots, but because it builds the brains that power them. The company's ambitious goal: create the equivalent of ChatGPT, but for the physical world.

The Dream Team Behind π

Physical Intelligence — stylized as π — was founded in early 2024 by a group that reads like a who's who of robotics AI research.

Karol Hausman, the CEO, was previously a Staff Research Scientist at Google DeepMind, where he developed advanced manipulation and locomotion algorithms. Sergey Levine, serving as Chief Scientist, is a UC Berkeley professor who pioneered deep reinforcement learning for robotics — his lab is arguably the most influential in the field. Chelsea Finn, an assistant professor at Stanford, is one of the most-cited researchers in meta-learning and previously worked at Google. Brian Ichter, another ex-DeepMind researcher, specialized in robotic planning, control, and trajectory optimization.

Rounding out the founding team are Lachy Groom, a former Stripe executive who brings Silicon Valley business acumen, and Adnan Esmail. It's a rare combination: world-class academic researchers paired with proven startup operators.

As Levine has put it simply: "Think of it like ChatGPT, but for robots."

A Funding Trajectory That Defies Gravity

Physical Intelligence's fundraising history tells the story of a market moving at breakneck speed:

• Seed Round (March 2024): $70 million
• Series A (November 2024): $400 million at a $2 billion valuation — investors included Jeff Bezos personally, OpenAI, Thrive Capital, Lux Capital, and Bond Capital
• Series B (November 2025): $600 million at a $5.6 billion valuation — led by CapitalG (Alphabet's growth fund), with participation from Sequoia Capital, NVIDIA (NVentures), Redpoint Ventures, and T. Rowe Price
• Current Discussions (March 2026): $1 billion at $11 billion+ valuation — Founders Fund and Lightspeed Venture Partners reportedly joining, with Thrive Capital and Lux Capital expected to follow on

The valuation trajectory — $2B → $5.6B → $11B+ in under 18 months — is extraordinary even by AI startup standards. The investor roster is equally telling: when Jeff Bezos, OpenAI, NVIDIA, and Alphabet's investment arm all back the same robotics company, it signals broad consensus that this team is onto something real.

The Technology: From π0 to π0.6

At the heart of Physical Intelligence's value proposition is its Vision-Language-Action (VLA) model — a new class of AI that can see the world, understand natural language instructions, and generate physical robot actions.

π0: The Foundation (October 2024)

The company's first major model, π0, builds on Google's PaliGemma — a 3-billion-parameter vision-language model — augmented with an additional 300 million parameters dedicated to robot control. Trained on data from 7 robotic platforms across 68 unique tasks, π0 introduced a novel "flow matching" architecture enabling high-frequency control at up to 50 times per second. This allows robots to perform dexterous tasks like folding laundry, preparing coffee, and assembling boxes.

π0.5: Open-World Generalization (2025)

The breakthrough with π0.5 was open-world generalization — the ability to operate in completely new environments. Physical Intelligence tested the model in three rental homes across San Francisco, demonstrating that a mobile manipulator could clean up kitchens and bedrooms it had never seen before. This was a critical milestone: moving from controlled lab settings to messy, unpredictable real-world spaces.

π0.6: Learning from Experience (November 2025)

The latest model introduced RECAP, a reinforcement learning approach where robots learn through a three-stage process: demonstration, coaching corrections, and autonomous experience. The result was a 2x improvement in throughput on tasks like espresso machine filter insertion, folding unseen laundry, and cardboard box assembly.

Critically, Physical Intelligence has open-sourced the π0 model's code and weights, inviting the developer community to fine-tune it for their own robots and tasks. This is a strategic play to establish π0 as the standard platform for robotics AI — the Linux of robot brains.

The Competitive Landscape: A Crowded but Stratified Market

Physical Intelligence operates in what has become 2026's most competitive investment category. The AI robotics sector is projected to attract over $20 billion in funding this year alone, and the humanoid robot market is expected to grow from $2.92 billion in 2025 to $15.26 billion by 2030.

But the competitive landscape is more nuanced than it appears. Physical Intelligence occupies a distinct niche:

Figure AI — valued at $48 billion after its Series D — is the current market leader in humanoid robotics hardware. Its Amazon warehouse deployment (20,000 units, ramping from 1,200/month to 5,000/month by Q4 2026) represents the first major commercial validation of humanoid robots at scale.

Skild AI raised $1.4 billion in a Series C in January 2026, positioning itself as Physical Intelligence's most direct competitor in the robot foundation model space, with a valuation exceeding $14 billion.

Tesla's Optimus benefits from virtually unlimited internal R&D funding and Tesla's manufacturing expertise, targeting $20,000-$30,000 per unit and 1 million units/year by 2030.

Boston Dynamics, under Hyundai's ownership, announced a strategic partnership with Google DeepMind at CES 2026 to integrate Gemini Robotics models into its Atlas platform.

Physical Intelligence's differentiation is clear: it's not building robots — it's building the universal operating system for robots. While Figure AI, Tesla, and Boston Dynamics compete on hardware, Physical Intelligence aims to be the AI layer that any robot maker can adopt. It's the Android strategy applied to robotics.

What Will They Do With $1 Billion?

If the round closes as reported, Physical Intelligence will have raised over $2 billion in total — a war chest that enables several strategic moves.

Compute infrastructure is likely the largest line item. Training next-generation VLA models requires massive GPU clusters, and the company's strategic relationship with NVIDIA suggests significant investment in compute capacity. The jump from π0.6 to whatever comes next will demand exponentially more training resources.

Commercialization is the next frontier. Physical Intelligence has operated primarily as a research organization, but with $2 billion in the bank, the pressure to demonstrate commercial traction intensifies. Expect partnerships with major robot manufacturers, logistics companies, and potentially consumer electronics firms looking to add intelligence to their hardware.

Talent acquisition remains critical. The AI robotics talent pool is shallow, and competition for top researchers is fierce among Physical Intelligence, Google DeepMind, Tesla, Figure AI, and a growing number of well-funded startups. A fresh billion-dollar round provides the compensation packages needed to recruit and retain the best.

Why Investors Are Betting Big

The bull case for Physical Intelligence rests on platform economics. If the company successfully establishes its VLA models as the standard AI layer for robotics, it becomes the indispensable software provider to an entire industry — collecting licensing fees or API revenue from every robot maker that adopts its technology.

The analogy investors keep returning to is the smartphone ecosystem. Just as Google's Android became the default operating system for the majority of the world's smartphones, Physical Intelligence could become the default intelligence layer for the world's robots. The total addressable market, if robotics achieves even a fraction of smartphone-scale adoption, would be measured in hundreds of billions of dollars.

The founding team's academic credentials create a formidable moat. Levine and Finn are among the most-cited researchers in their respective fields, and their lab alumni networks give Physical Intelligence a structural advantage in recruiting that is nearly impossible to replicate.

There's also a timing argument. Apple, Meta, and Google DeepMind are all entering the robotics space, signaling that Big Tech sees embodied AI as the next major platform. Physical Intelligence, as an independent startup with a two-year head start on foundation models, is positioned as a potential acquisition target — or a potential platform company in its own right.

What to Watch

Physical Intelligence's reported $11 billion valuation negotiation marks a symbolic moment: the AI robotics industry's transition from research curiosity to serious commercial contender. The key milestones to watch are the final terms of this round, the company's first major commercial partnerships, and the performance of its next-generation models. With $2 billion in cumulative funding, a founding team that includes some of the most accomplished robotics researchers alive, and a market that's growing faster than anyone predicted, Physical Intelligence has all the ingredients to become the defining company of the robotics AI era. Whether it can execute on that promise — and justify an $11 billion valuation before generating meaningful revenue — is the billion-dollar question.