The Most Expensive Robotics Research Lab Just Got More Expensive
Physical Intelligence (PI), the San Francisco-based embodied AI startup, is reportedly in advanced discussions to raise ~$1 billion in new funding at a valuation exceeding $11 billion—nearly doubling its $5.6 billion post-money valuation from just four months ago, according to Bloomberg.
If the round closes, it would mark one of the fastest valuation escalations in robotics history—and cement PI’s position as the sector’s most aggressive “research-first” bet.
The Thesis: “ChatGPT for Robots”—But Without the Revenue Clock
Founded in 2024 by a dream team of robotics researchers—Karol Hausman (ex-Google DeepMind), Sergey Levine (UC Berkeley), Chelsea Finn (Stanford), Brian Ichter (ex-Google Research), and Adnan Esmail—PI is building general-purpose foundation models for physical intelligence.
Its flagship model, π0 (pi-zero), launched in October 2024 and open-sourced in February 2025, uses a vision-language-action (VLA) to generalize across tasks: laundry, coffee-making, box assembly, vegetable peeling
Critically, PI does not build robots.
It builds the software brain that can run on any robot.
“If someone builds a new hardware platform tomorrow, they won’t need to start data collection from scratch,” Vuong explains.
Physical Intelligence Funding: Who Is Betting $1B?
PI has raised over $1 billion across three rounds in under two years. The reported new round would include:
| Investor Type | Participants |
|---|---|
| New Leads | Founders Fund (Peter Thiel), Lightspeed Venture Partners |
| Returning Backers | Thrive Capital, Lux Capital, Khosla Ventures, Sequoia Capital |
| Strategic/Corporate | NVIDIA (NVentures), Jeff Bezos (via Bezos Expeditions), Index Ventures, T. Rowe Price |
Notably, PI has no commercialization timeline—an unusual posture that its investors reportedly accept.
Co-founder Lachy Groom (ex-Stripe, early angel in Figma/Notion/Ramp) is explicit:
“I don’t give investors answers on commercialization. That’s sort of a weird thing, that people tolerate that.”
Instead, capital flows almost entirely to compute and data collection:
- ~80 employees (mostly researchers)
- Low-cost robotic arms (~$3,500/unit) for real-world data gathering
- Partnerships with warehouses, test kitchens, and homes for diverse environment data
They are not the only one tho.
PI’s strategy stands in stark contrast to competitors like Skild AI, which just raised $1.4B at $14B with $30M in early revenue from commercial deployments in security, logistics, and manufacturing.
| Dimension | Physical Intelligence | Skild AI |
|---|---|---|
| Core Focus | General-purpose foundation models | Omni-bodied commercial deployments |
| Revenue | None disclosed | ~$30M (2025) |
| Hardware | Agnostic (uses cheap arms for data) | Integrated (Skild Brain + partner hardware) |
| Data Strategy | Diverse real-world collection + simulation | Real-world deployments + simulation |
| Investor Pitch | “Build the best brain first” | “Deploy now, learn faster” |
📌 Investment Takeaway: The Optionality Trade
For institutional investors, PI represents a high-conviction optionality bet:
The key metric to watch:
Not quarterly revenue—but model generalization benchmarks and partner adoption signals.
If π1 (or π2) demonstrates zero-shot transfer across 10+ robot platforms with human-level task success, the $11B valuation may prove conservative.
If not, PI risks becoming the most expensive robotics research project in history.
Remember
Physical Intelligence isn’t trying to ship robots.
It builds the software brain that can run on any robot.
