What Dexterity does
Dexterity is a physical AI company building specialized robotic foundation models and full-stack software for warehouse and supply-chain robots. Where humanoid-robot startups chase a single hardware platform, Dexterity layers its AI onto industrial-grade arms and gantries to perform high-value, repetitive tasks: truck loading and unloading, palletizing and depalletizing, parcel sorting, and goods-to-person picking. Each task is handled by a purpose-trained model, with the whole stack coordinated by Dexterity's robot operating layer.
The company brands its production deployments as Robots-as-a-Service, with FedEx, UPS, and other Fortune 50 logistics operators running Dexterity robots in live distribution centers. By late 2025 the company crossed 100 million autonomous actions in production.
Who it's for
Dexterity targets 3PLs, parcel carriers, retailers, and manufacturers with high-volume distribution operations. Buyers are typically VPs of operations, supply chain, or automation who want robots that pay back through measurable throughput rather than science projects.
Pricing
Dexterity sells primarily as Robots-as-a-Service — a per-robot, per-hour or per-action recurring fee that bundles hardware, software, and ongoing model updates. Pricing is enterprise.
Team & funding
Dexterity was founded in 2017 by Samir Menon (CEO), a Stanford PhD whose research focused on robotic control and simulation, in Redwood City, California. The company has raised approximately $300M total. Its latest round, a $95M financing in March 2025 led by Lightspeed Venture Partners with Sumitomo Corp, came at a $1.65B valuation. Earlier investors include Kleiner Perkins, Obvious Ventures, and Cathay Innovation.
Position vs competitors
Dexterity competes with humanoid-robot OEMs like Figure and Apptronik, horizontal robot-brain plays like Skild AI and Physical Intelligence, and warehouse-automation vendors like Symbotic and Berkshire Grey. Its differentiator is shipping production specialized models on industrial arms today rather than waiting on humanoid-platform maturity.