Dexterity is an enterprise Physical AI platform that gives robots human-like dexterity for autonomous loading, palletizing, and logistics tasks. See pricing,…
Category
Internet of Things (IoT) & Home Automation
Pricing
Not publicly disclosed — Dexterity does not list plans or self-serve pricing; enterprises must contact the company (contact@dexterity.ai) to discuss custom deployment terms.
Verified
Not yet
Last updated
August 6, 2026
EnterpriseAutomationAPIAIMachine Learning
Dexterity is an enterprise Physical AI company, founded in 2017 by Stanford roboticist Samir Menon and five co-founding engineers, headquartered in Redwood City, California. The company builds AI systems that give industrial robots "human-like dexterity" so they can perceive, reason about, and act autonomously in unstructured real-world environments such as warehouses, distribution centers, and manufacturing lines. Its core technology is Foresight, a world model for Physical AI trained on over 100 million autonomous production actions, coordinated by a fleet of 68+ specialized "Autonomous Skill Agents" that handle perception, motion planning, six-axis force/tactile control, and dual-arm task orchestration. Its IRIS layer is a hardware-agnostic abstraction API that decouples the software from specific robotic arms or mobile platforms, with automatic camera calibration and microsecond-accuracy torque control. Dexterity's systems are used in production (not pilot) deployments for tasks like trailer and aircraft loading/unloading, palletizing, depalletizing, and parcel singulation. Reported customers and partners include FedEx, Maersk, VFC, Sumitomo Corporation, and Sagawa Express, and the company has expanded operations into Japan. It is backed by investors including Kleiner Perkins and Lightspeed Venture Partners and has been covered by Forbes, Bloomberg, and The Wall Street Journal. In 2026 Dexterity opened a Foresight API to outside developers through a public API Challenge program.
Key Features
Foresight world model — A Physical AI world model trained on over 100 million autonomous production actions, letting robots perceive 3D environments, predict outcomes, and act autonomously.
68+ Autonomous Skill Agents — A coordinated fleet of specialized agents handling perception, motion planning with collision avoidance, and real-time task allocation for complex manipulation tasks.
IRIS hardware abstraction layer — A hardware-agnostic universal API that decouples the software from specific robotic arms or mobile platforms, with automatic camera calibration and hardware feature discovery.
Six-axis force and tactile control — Force control with six-axis tactile sensing and microsecond-accuracy torque control for precise physical manipulation.
Dual-arm task orchestration — Coordinates dual-arm robots with task orchestration and parallelism for operations like palletizing, depalletizing, and singulation.
Continuous safety monitoring — Ongoing safety monitoring described by Dexterity as providing software-transaction guarantees on physical robot behavior.
Pros & Cons
Pros
Production-proven at scale, with a reported 100M+ autonomous actions completed and zero safety incidents claimed to date
Hardware-agnostic IRIS abstraction layer lets the same software run across different robotic arms and mobile platforms
Backed by well-known investors (Kleiner Perkins, Lightspeed) and used by large enterprise clients such as FedEx and Sumitomo Corp
Focused on real production deployment rather than demos or pilots, per company messaging
Cons
No public pricing — enterprises must contact sales directly, with no listed plans or self-serve pricing on the site
Built for large-scale enterprise logistics and manufacturing operations, not a fit for small businesses or individual use
Requires physical robotic hardware and integration work, not a pure self-serve software purchase
Limited public technical documentation outside of the invite-based Foresight API Challenge program