Physical AI

Market Context

Physical AI — the convergence of foundation models, robotics, and autonomous systems operating in the real world — is experiencing one of the most concentrated capital surges in venture history, with $38.6B deployed across 12 deals in the past 28 days alone. The market has moved decisively beyond software-only AI: investors including BlackRock, JPMorgan, and Goldman Sachs are writing mega-checks into companies that compress physical engineering cycles, deploy humanoid robots, and build hierarchical vision-language-action (VLA) systems. Research from Google DeepMind is simultaneously providing the architectural blueprints that will define which robot policy designs win at scale.

Investment Activity

• Project Prometheus raised a $12B growth round at a $41B valuation, led by JPMorgan, BlackRock, and Goldman Sachs, with participation from DST Global and Arch Venture Partners.
• Figure raised a $1.4B Series C with investors including Tether, Qualcomm, Amazon, Nvidia, and Bosch.
• A $27M Series A was raised with Gokul Rajaram as lead investor.

Key Players

• Project Prometheus: Jeff Bezos-backed physical AI company targeting a $10B fundraise at a $38B valuation, building tools to compress industrial engineering cycle times up to 10x faster for aerospace, medical devices, and consumer electronics.
• Physical Intelligence: Leading VLA robotics startup whose π0.5 and π0.7 hierarchical systems are cited in Google DeepMind research as validation that smaller high-level VLMs can drive strong hierarchical robot performance.
• Figure: Humanoid robotics company whose Helix hierarchical VLA system is cited in Google DeepMind's systematic study as a production deployment using fixed-timer termination, attracting a $1.4B Series C from strategic investors including Nvidia and Amazon.
• Waymo: Alphabet's autonomous vehicle unit cited as the benchmark success story for slow-and-steady physical AI deployment, with commercial robotaxi operations in San Francisco validating the long-horizon investment thesis.
• Galaxea: Robotics company whose G0 dual-system hierarchical VLA model is cited in Google DeepMind research (arXiv 2606.10267) as a motivating production example of divergent hierarchical VLA implementations.

Market Signals

• Google DeepMind published the first systematic head-to-head comparison of hierarchical VLA design choices (arXiv 2606.10267), finding reasoning capability matters more than model scale — a signal that mid-sized models can compete with frontier-scale systems in robot control.
• KKR, Nvidia, and the Kuwait Investment Authority launched Helix Digital Infrastructure, a new data center group with $10B in capital commitments, signaling that physical AI compute infrastructure is being institutionalized at sovereign scale.
• JPMorgan, BlackRock, and Goldman Sachs co-investing in Project Prometheus marks a shift from VC-led to institutional-capital-led funding in physical AI, compressing the timeline to mega-round liquidity events.
• Nvidia's GR00T N1 hierarchical VLA system is cited alongside Figure's Helix and Galaxea's G0 as production generalist humanoid robot control systems, confirming Nvidia is competing as a platform player — not just a compute supplier — in physical AI.
• Geographic momentum is concentrated in the United States (Project Prometheus in SF, Figure, Waymo, Physical Intelligence, 1X Technologies in Hayward CA) with China-linked players (Galaxea, Horizon Robotics, Astribot) accelerating in parallel on foundation model robotics research.