Dexterity

Dexterity is a physical-AI company founded in 2017 and headquartered in Redwood City, California, that builds robots capable of perceiving and manipulating the endless variety of objects moving through warehouses and factories. Its systems load and unload trucks, palletize mixed cases, and sort parcels — tasks that defeat traditional fixed automation because no two boxes are quite alike.

The load-bearing technology is learned manipulation. Dexterity's robots use AI perception to see cluttered, variable scenes and learned grasp planning to decide how to pick and place each item, adapting on the fly rather than following a pre-programmed motion. The company raised roughly ninety-five million dollars in 2025 at a valuation near 1.65 billion dollars, and expanded a manipulation partnership with Kawasaki Robotics in 2026.

The honest framing: Dexterity is a genuine physical-AI bet where perception and learned manipulation are unambiguously the differentiator. Like all manipulation robotics, its progress is measured task by task — the hard part is reliability across the long tail of odd-shaped, deformable, and unexpected items.