Dexterity (company)
Last reviewed
Sources
20 citations
Review status
Source-backed
Revision
v2 · 2,093 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
20 citations
Review status
Source-backed
Revision
v2 · 2,093 words
Add missing citations, update stale details, or suggest a clearer explanation.
Dexterity (Dexterity, Inc., often styled Dexterity AI) is an American robotics company based in Redwood City, California, that builds AI-powered robotic systems for warehouses, logistics, and supply chains. Founded in 2017 by Stanford robotics PhD Samir Menon, who is the company's chief executive, Dexterity develops what it calls "Physical AI": software that gives industrial robot arms a sense of touch, force control, and visual reasoning so they can palletize and depalletize, pick, pack, sort, and load and unload trucks. The company emerged from stealth in July 2020 and reached unicorn status in 2021. A March 2025 funding round of $95 million valued it at roughly $1.65 billion. Its robots have been tested or deployed by global logistics firms including FedEx, UPS, GXO Logistics, and Japan's Sagawa Express, and it runs a joint venture with Sumitomo Corporation to sell its systems in Japan.
Despite the name, the company should not be confused with the general concept of manual dexterity, nor with unrelated software products that share the "Dexterity" name (such as Microsoft's Great Plains Dexterity development environment). Dexterity AI is specifically a physical AI and warehouse robot company.
Dexterity was founded in 2017 in Redwood City, California, by Samir Menon, who serves as founder and chief executive officer. Menon earned a PhD in robotics from Stanford University; his doctoral research developed a control-theory framework describing how the human brain coordinates and controls the body, work that informed the company's approach to giving robots fine motor control and a sense of touch. Founding engineers and co-founders named by the company include Robert Sun, Kevin Chavez, Ben Varkey Benjamin, Talbot Morris-Downing, and Cuthbert Sun.
The company operated in stealth for roughly three years before its public launch. Its central thesis, in the founders' framing, was that the next major advance in artificial intelligence would not be another chatbot but rather giving machines the intelligence to operate in unstructured physical environments. Dexterity built a full-stack system spanning hardware-agnostic robot control software, computer vision, and tactile sensing, rather than manufacturing robot arms itself; early systems used arms from suppliers such as Kawasaki Heavy Industries.
Dexterity has raised approximately $291 million across its disclosed rounds. The company emerged from stealth in July 2020 with a Series A; an October 2021 Series B made it a unicorn at a roughly $1.4 billion valuation; and a March 2025 venture round of $95 million raised the post-money valuation to about $1.65 billion, bringing total funding to nearly $300 million.
| Round | Date | Amount | Lead / notable investors | Valuation |
|---|---|---|---|---|
| Series A | July 2020 | $56.2 million | Kleiner Perkins, Lightspeed Venture Partners, Obvious Ventures, B37 Ventures, Presidio Ventures (Sumitomo), Blackhorn Ventures, Liquid 2 Ventures, Stanford StartX | Not disclosed |
| Series B | October 2021 | $140 million | Lightspeed Venture Partners and Kleiner Perkins, with Obvious Ventures, B37 Ventures, Presidio Ventures | ~$1.4 billion (unicorn) |
| Venture round | March 2025 | $95 million | Lightspeed Venture Partners and Sumitomo Corporation | ~$1.65 billion |
Lightspeed Venture Partners and Kleiner Perkins have backed the company since its early rounds. Sumitomo Corporation first invested in 2020 through its U.S. corporate venture arm, Presidio Ventures, and went on to become a strategic partner and co-lead of the 2025 round.
Dexterity describes its core technology as Physical AI, broadly meaning the application of AI techniques to systems that physically interact with the real world. Rather than relying on a single large model, the company uses what it has described as an "AI of AIs" or orchestration approach: hundreds of specialized AI models, each focused on a narrow task such as perceiving objects, planning how to stack boxes, or controlling gripping force, coordinated together in real time. Menon has summarized the goal as robots that can touch and recognize objects, stay aware of and respond to their surroundings, and move gracefully while adjusting on the fly.
Distinguishing technical elements the company has highlighted include:
The company's commercial software is marketed simply as the Dexterity Platform (it offers task-specific applications such as palletizing and depalletizing software), and it positions its systems within the broader fields of robot manipulation, embodied AI, and robot foundation models.
Dexterity sells its robots largely under a Robots-as-a-Service (RaaS) model and targets logistics tasks that are repetitive, physically demanding, or hazardous, including industrial palletizing, depalletizing, singulation and induction of parcels, kitting, order fulfillment, and truck and container loading and unloading.
DexR is Dexterity's mobile trailer-loading robot, unveiled in a collaboration with FedEx at a joint "Unlock the Dock" event in San Francisco on September 26, 2023. DexR navigates autonomously to the back of a trailer and uses two robotic arms to load boxes from a powered conveyor into the truck, picking and packing simultaneously to improve throughput. Its capabilities include generative wall planning, a built-in sense of touch, machine-learning-based pack improvement, and integrated motion planning. Reported specifications include a payload around 60 kg and a reach of more than five meters, with operation in temperatures from 32 to 122 degrees Fahrenheit. As of the 2023 to 2025 period, FedEx was conducting ongoing testing with a stated goal of eventual commercial deployment. Dexterity has also tested trailer-loading technology with Sagawa Express and GXO Logistics.
Mech is a dual-armed mobile manipulation robot (MMR) that Dexterity launched on March 18, 2025, and markets as the world's first "industrial superhumanoid." It is important to be precise about what Mech is: it is not a walking, bipedal humanoid. Mech mounts two industrial robot arms on a wheeled rover base with four independently steerable wheels, allowing it to drive, spin, and strafe between workstations across a warehouse or industrial site and then perform heavy lifting, loading, palletizing, depalletizing, picking, and packing. Dexterity frames it as combining human-like adaptability with superhuman strength rather than imitating human form.
Reported and company-stated specifications for Mech include:
| Attribute | Specification |
|---|---|
| Configuration | Two industrial arms on a four-wheel steerable rover base |
| Lift capacity | Around 60 to 65 lb (29 kg) per arm; roughly 130 lb (59 kg) combined |
| Reach / height | Arms span about 5.4 m; can place boxes up to about 8 ft (2.4 m) high |
| Sensing | Machine vision (up to 16 onboard cameras), force sensing, AI-driven sense of touch and pressure with ~4 kHz torque control |
| Compute | Onboard AI supercomputer running Dexterity's Physical AI |
| Autonomy | Company cites dozens of "skill agents" orchestrated in real time |
| Operating range | About 0 to 50 degrees C (32 to 122 degrees F) |
| Supervision | One associate can monitor up to roughly 10 units |
| Arm supplier | Kawasaki; designed and assembled in California |
Mech became available to industrial and enterprise customers initially for truck loading, with additional software applications planned through 2025 to extend its tasks.
FedEx has been a flagship collaborator. The two companies co-developed and publicly demonstrated the DexR trailer-loading robot in September 2023, and FedEx has continued testing the technology in its operations. FedEx executive Rebecca Yeung and Menon have both publicly framed the partnership around automating the most difficult and physically taxing tasks in package operations.
Sumitomo Corporation has been involved with Dexterity since investing in 2020 via Presidio Ventures. In October 2022 the two signed a distributorship agreement under which Sumitomo became the exclusive distributor of Dexterity's products and services in Japan, with a stated goal of deploying 1,500 robots across Japanese warehouses by 2026. In June 2024 they deepened the relationship by establishing a joint venture, Dexterity-SC Japan, based in Tokyo and led by Shotaro Oyama. The joint venture sells, markets, localizes, and arranges financing (via Sumitomo Mitsui Finance and Leasing) for Dexterity's AI-powered robotic solutions in the Japanese market, focused on truck loading and unloading, palletizing, and depalletizing for labor-short logistics operations. Sumitomo co-led Dexterity's March 2025 funding round.
On May 29, 2025, Dexterity announced a partnership with contract manufacturer Sanmina Corporation to scale production of the Mech robot. Under the agreement, Mech is assembled and tested at Sanmina's advanced manufacturing facilities in California, which Dexterity has highlighted as a commitment to domestic production. The companies framed the partnership as building manufacturing capacity to meet demand across parcel, ground logistics, e-commerce, retail, and air cargo.
Beyond FedEx and the Japanese deployments, Dexterity's customers and named partners over time have included UPS, GXO Logistics, Kawasaki Heavy Industries, Maersk, and various global food and consumer-goods manufacturers. The company reported scaling from about 10 million autonomous production actions in 2023 to about 100 million by 2025.
Dexterity sits among a cohort of well-funded startups applying AI to industrial and warehouse robotics, a field that drew heavy venture interest in the mid-2020s alongside covariant, Pickle Robot, and Anyware Robotics, and adjacent to humanoid developers such as Figure AI, Agility Robotics, and Apptronik. It competes in trailer unloading with systems like Boston Dynamics' Stretch. Dexterity was named to The Robot Report's RBR50 Robotics Innovation Award list in 2024 for its work stacking AI and robotics for truck loading. Coverage has noted that, unlike many peers chasing general-purpose humanoids, Dexterity has concentrated on stationary and wheeled industrial manipulators aimed at concrete logistics throughput, paired with a strategy of partnering with manufacturers (Sanmina) and distributors (Sumitomo) rather than building everything in-house.