Who are Physical Intelligence's decision-makers?

The company is led by CEO Karol Hausman, a former Staff Research Scientist at Google DeepMind and adjunct professor at Stanford University. Hausman holds a PhD from the University of Southern California, built his reputation on robot manipulation learning and cross-embodiment policies, and also serves as Secretary and CFO per California corporate filings — reflecting the company's still-lean management structure.

Chief Scientist Sergey Levine is one of the most-cited researchers in robot learning globally. His UC Berkeley lab — where much of the academic grounding for Physical Intelligence's models was developed — remains a primary pipeline for research talent. Chelsea Finn (Stanford), another co-founder, brings expertise in meta-learning (her MAML algorithm is foundational to cross-task generalization) and sim-to-real transfer; she helps define the research agenda while on leave from her Stanford professorship. VP Engineering Brian Ichter, also from Google DeepMind and Google Brain, leads the engineering organization responsible for training infrastructure and model deployment systems.

COO Lachy Groom is the commercial operator of the founding team. His background as a Stripe product leader — and angel investor in Figma, Notion, and Ramp — makes him unusual in a company otherwise staffed with academic researchers. He owns fundraising, business development, and the partnership program, and is the registered agent on Physical Intelligence's California corporate filings. Two additional co-founders add depth: Adnan Esmail (former SVP of Engineering at Anduril Industries and senior Tesla Autopilot engineer; MIT-trained in Mechanical Engineering, Physics, and EECS) brings hardware systems expertise, while Quan Vuong (Google DeepMind, UC San Diego) contributes deep RL and data pipeline experience.

Physical Intelligence does not have a traditional procurement function at its current scale (~200 employees). All significant vendor and partnership decisions flow through the co-founder level. For technical tooling — compute, MLOps infrastructure, robotics hardware, data annotation platforms — the decision-maker is VP Engineering Brian Ichter, with input from the research leads (Levine, Finn, Hausman). Given the company's deep research culture, technical evaluations are rigorous and engineer-led; pure business-case presentations without a technical co-sell will not land.

For commercial partnerships — robotics companies wanting to license or integrate Physical Intelligence's models — the primary engagement point is COO Lachy Groom via partnership@physicalintelligence.company. The company published a formal partner blog post describing its collaboration model, which suggests the business development process is becoming more structured. Final partnership decisions are made jointly by Groom and Hausman. The fastest path to a meeting is through a warm introduction from a shared investor (Thrive, Sequoia, CapitalG, Lux) or through a technical champion who already works with openpi on GitHub.

As the company scales toward commercial deployments with partners like Weave Robotics and Ultra, a more formal sales and customer success function will likely emerge under Groom's remit. The introduction of a dedicated CRO or VP of Sales would be the clearest signal that Physical Intelligence has transitioned from research-first to revenue-first.

With ~200 employees (up from approximately 80 in early 2025), Physical Intelligence remains a flat, founder-led organization. Research, engineering, and commercial operations are not yet siloed into separate business units. Hiring has been deliberately selective — the company targets elite researchers with published records in policy learning, reinforcement learning, and robotic manipulation, and has grown headcount roughly 2.5x in approximately 12 months.

The seven-person founding team covers almost all critical functions: research vision (Levine, Finn, Hausman), engineering execution (Ichter), hardware systems (Esmail), data and RL engineering (Vuong), and commercial operations (Groom). This breadth is intentional — the founders wanted to avoid the gap between research and productization that has plagued other robotics startups. As revenue becomes more material, expect headcount growth to concentrate in commercial engineering (customer integration, partner support), ML infrastructure (training and serving at scale), and business development.

Watch for VP-level hires in business development, customer success, and enterprise sales as leading indicators of Physical Intelligence's transition from a research-first lab to a commercially oriented software company. Those hires will likely report to Groom and will constitute the first layer of management between the founding team and the growing workforce.