Dexterity
Integrator / OEM
Redwood City physical-AI company selling full-stack warehouse robots-as-a-service for truck loading, palletizing and parcel handling.
Dexterity builds full-stack 'physical AI' for warehouse manipulation: dual-armed mobile robots (Mech) and a world-model software layer (Foresight) that give industrial arms human-like dexterity. Its systems perform truck and trailer loading/unloading, palletizing and depalletizing, parcel singulation and goods-to-person picking, deployed in existing infrastructure without major reconfiguration and sold as robots-as-a-service with a performance guarantee. Dexterity reports 100M+ autonomous production decisions with zero safety incidents and sub-400 ms decision speed, and works with FedEx, UPS, Sagawa Express and Maersk. (Company-reported facts.)
Research mapping
Physical AI / Autonomy Software, Fleet Platforms & End Markets (L10) / Integrator / OEM
Evidence boundary
Original analysis and independent corroboration remain under development. Commercial scale and durable technical advantage remain under review.
What makes it defensible
An 'AI of AIs' architecture coordinating hundreds of manipulation models plus a Foresight world model - full-stack intelligence that generalizes across messy, unstructured warehouse tasks.
Position in the stack: Vertical physical-AI integrator: a software brain plus robot cells sold as a service, layering intelligence onto industrial arms rather than selling components.
Who buys it
Named customers: FedEx, UPS, Sagawa Express, Maersk
Founding team
Samir Menon — Founder & CEO — Ph.D. in Computer Science, Stanford (robotic control and simulation of the human musculoskeletal system); founded Dexterity in 2017 in Redwood City with a team of Stanford roboticists.
SourceFinancing detail
- Last round
- Late Stage
Investors: Lightspeed Venture Partners, Sumitomo Corporation, Kleiner Perkins
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.
Evidence & media
-
Introducing Foresight — a physical-AI world model
What this source supports: First-party: Foresight world model trained on 100M+ autonomous production actions; Foresight API Challenge launched Mar 2026.
What it does not establish: First-party; 100M-action count and safety record self-reported.
Source -
Foresight world model / 4D packing agent (trade press)
What this source supports: Trade press: Foresight makes each placement decision in <400 ms, proven across six applications, 4 robot types, 5 hand types.
What it does not establish: Restates first-party release; no independent benchmark.
Source -
Dexterity's Mech — industrial super-humanoid (Fox News)
What this source supports: Mass-market coverage of Mech specs (large reach, ~130 lb lift, up to 16 cameras) and physical-AI positioning.
What it does not establish: Consumer-tech coverage; specs are vendor-supplied.
Source -
Kawasaki Robotics partners with Dexterity to advance warehouse logistics
What this source supports: RL030N 8-DoF arm platform production-ready for Dexterity Mech; sourced to joint media release.
Source