Sensei Robotics
Data & simulation
Y Combinator-backed 'Scale AI for robotics data' — a marketplace of human operators collecting demonstration data on low-cost sensorised hardware.
Sensei Robotics runs a robot training-data engine: a marketplace of paid human operators ('Senseis') who collect in-the-wild human-demonstration data used to train robot manipulation policies. Operators use a low-cost, portable sensorised exoskeleton arm (under ~$300) that captures joint angle, vision and inertial signals while they perform real tasks — clothes folding, bin sorting, dish loading — for robotics customers who submit bespoke data requests. Customers get labelled demonstration data at roughly a tenth of the cost and twice the speed of conventional teleoperation, according to the company's positioning. The pitch is explicitly 'Scale AI for robotics data': the bottleneck is not compute or models but the supply of high-quality real-world human demonstration data, and Sensei productises that supply.
Research mapping
Physical AI / AI Factory & Cloud Training Infrastructure (L1) / Data & simulation
Evidence boundary
Original analysis and independent corroboration remain under development. Commercial scale and durable technical advantage remain under review.
What makes it defensible
Cheap (<$300) sensorised capture rig plus a trained operator network — the data-supply flywheel, not the model, is the scarce input in robot learning.
Position in the stack: Data-layer supplier sitting upstream of every robot foundation-model lab and OEM: it sells the demonstration episodes that policies are trained on, so it sits on the input side of the whole manipulation stack.
Who buys it
Founding team
Anubhav Guha — Co-founder & CEO — MIT-trained robotics engineer; left an MIT PhD in robotics to start Sensei; previously worked on AI for autonomous fighter jets at Aurora Flight Sciences on a DARPA programme
John Piotti — Co-founder & CTO — MIT engineer; led reinforcement-learning efforts at Aurora Flight Sciences on the same DARPA autonomous-fighter programme
SourceFinancing detail
- Last round
- Pre-Seed
Investors: Y Combinator, Network.VC
Mapped funds and investors
Direct relationships supported by published evidence.
Reported holdings as-of
No sourced relationships mapped yet.
Disclosed investment history
No sourced relationships mapped yet.
Showing up to four relationships in each group. ETF weights describe the fund’s portfolio, not ownership of this company. Private investment history does not establish current ownership.