November 18, 2026:


Genesis AI announced Tuesday that it raised a $500 million Series A at a $3.5 billion post-money valuation, but the Eno robot’s hands last only a few weeks of regular work, and TechCrunch’s reporting on Eno says the demonstration it saw did not use the robot itself. Premji Invest, Khosla Ventures and Eclipse co-led the round, which funds data collection, model training and production of the wheeled, two-armed robot. Buyers weighing Eno should read the funding figure as a bet on a data method, not as proof the hardware holds up.
Only 30 to 40 Eno units exist so far, according to TechCrunch. The Next Web reports that Genesis targets 2027 mass production, with the first customer deployments planned for the end of 2026. The company has not named most of its pilot customers in the sources reviewed, though The Next Web says LG CNS is testing Eno.
Genesis AI’s approach starts with people, not robots. More than 100 workers wear sensor gloves and headset cameras while doing ordinary tasks, and the recorded hand motion becomes training data for the company’s GENE-26.5 foundation model. The gloves carry tactile electronic skin, so the data includes touch as well as motion.
The economics are the pitch. Genesis says its gloves cost about $300 per pair, while collecting robot training data through teleoperation has been reported to cost far more per hour of operator time. Teleoperation means a person remotely drives a robot, so cost scales with every paid operator hour. A glove needs no robot in the loop, which breaks that link.
Genesis describes a 1:1:1 mapping between the glove, the human hand and the robot hand. Each point the glove senses corresponds to a joint on the human hand, which in turn corresponds to a joint on Eno’s hand. Eno’s hands have 20 actively driven degrees of freedom and are back-drivable, meaning the joints give way under outside force instead of resisting it.
The method targets a known weakness of imitation learning from demonstrations, where a robot learns a policy by copying examples. When a robot drifts into a situation the demonstrations never covered, errors compound, a problem researchers call distribution shift. A matched hand shape reduces the translation error between human and robot, but it does not remove the coverage problem.
Genesis also claims its gloves cut hardware cost by 100 times and that its approach improves data efficiency by up to 5 times. The sources reviewed do not describe how those figures were measured, so they remain company claims.
Robotics researcher Rodney Brooks has argued that dexterity, and touch sensing in particular, remains an unsolved problem for humanoid robots. Genesis’s glove design addresses the touch point directly.
Scale is the harder point. UC Berkeley’s Ken Goldberg has estimated a 100,000-year data gap between the physical data robots have and the text data that trained chatbots. One hundred glove wearers add data steadily, but nothing disclosed shows how many hours of data Genesis has collected or how well it transfers across tasks.
Simulation is the other half of the method. Genesis built its Genesis World simulator to test the model, and contact-rich tasks such as grasping remain the area where simulated physics most often diverges from real hands. TechCrunch did not identify which contact algorithm Genesis uses.
TechCrunch reports that Eno’s hands last a few weeks of regular work, a durability limit that matters for any buyer pricing replacement parts and downtime. Back-drivable, many-jointed hands are a long-standing cost and wear tradeoff in robotics, so the limit is an engineering constraint rather than an outlier.
The on-site demo used off-the-shelf arms, not Eno itself. That leaves the headline claim, skill learned from gloves and executed by Eno’s hands, without public independent proof.
Eno is a wheeled robot with two arms rather than a legged humanoid. Wheels avoid the balance problems of walking machines and put Eno in the same category as other mobile manipulators.
It could. French-American Genesis AI built its first units in China, according to Fast Company.
The location matters because of a July 28, 2026 FCC Public Notice, DA 26-786, which adds foreign-produced advanced robotic devices to the Covered List. The notice identifies covered devices by where they are produced, not by the maker’s nationality, and a Conditional Approval from the Department of War is the exemption path. The restrictions apply to new models. Whether Eno counts as a covered device depends on how the FCC applies that definition, and that question remains open.
Genesis has not said in the sources reviewed whether the China-built units are intended for U.S. customers or how its production plans relate to FCC authorization.
European sales face their own deadline. EU Machinery Regulation 2023/1230 applies from Jan. 20, 2027. Genesis has not published its certification plans.
Funding rivals are moving at the same scale. Agile Robots’ $500 million Series F came this month, and Neura reportedly raised up to $1.4 billion in a Series C. Genesis’s total disclosed funding is now $605 million after a $105 million round in July 2025. Investors are pricing hand dexterity and training data highly, and Tesla’s reported struggles with robot hands and worker training show why.
Whether the 100-plus glove wearers have consent, pay and data-retention terms has not been disclosed. Reporting has described Indian factory workers raising concerns about camera-based data collection for AI training. Tesla workers have reportedly resisted training Optimus robots. Neither report names Genesis, but employers considering glove programs should expect to answer the same questions.
Genesis AI raised a $500 million Series A at a $3.5 billion post-money valuation, announced Nov. 17. A post-money valuation includes the new money, so the pre-money valuation formula gives $3 billion. Total disclosed funding is $605 million.
Workers wear sensor gloves that record hand motion and touch while they perform tasks. Genesis maps each glove reading onto the human hand and then onto the robot hand, so the recorded motion can train a model that drives Eno’s hands.
That depends on FCC authorization. The FCC’s July 28, 2026 notice covers foreign-produced advanced robotic devices by production location rather than maker nationality, and Genesis has not said how it will address the rule for its first China-built units.
No independent verification was found. The 100x cost and up-to-5x data-efficiency figures are Genesis’s own claims, and the sources reviewed do not describe how they were measured.