By RWT Robotics Editorial Team Published
CONFIRMED
Genesis AI has announced a $500M Series A led by Khosla Ventures, Premji Invest and Eclipse, according to TechCrunch and Forbes, with Forbes reporting a $3.5B post-money valuation and listing Nvidia among the other investors. The money is meant to build general-purpose robots trained on glove-captured human dexterity for factories and warehouses. The hardware is still early: the company says it has built 30 to 40 Eno robots, and all of them are in pilots with a handful of customers it won’t name.
What we know
The round itself is solid. TechCrunch and Forbes both report that Genesis AI announced a $500M Series A, with Khosla Ventures, Premji Invest and Eclipse leading. Forbes puts the post-money valuation at $3.5B. Forbes also lists Nvidia among the additional investors.
That Nvidia line deserves a careful read. Nvidia’s participation rests on the Forbes investor list. We found no statement from Nvidia about the investment. “Nvidia-backed” is accurate as far as Forbes reports it. It does not tell you how large the check was or whether there is any technical partnership beyond the cap table.
On hardware, the numbers are modest and come from Genesis AI itself. The company told TechCrunch and Forbes that it has built 30 to 40 Eno robots and is piloting them with a small number of undisclosed customers. Forbes reports that industrial customers come first, specifically manufacturing, logistics and labs.
This is a fast climb. Genesis AI announced a $105M seed round in July 2025, so the company has raised roughly $605M in about 16 months. The co-founders are Zhou Xian and Thรฉophile Gervet.
The technical track record so far is one public model. In May 2026, TechCrunch reported that Genesis AI unveiled GENE-26.5, its first model, alongside a demo of robotic hands built with Wuji Tech.
What is rumored
Before the announcement, Bloomberg reported that Genesis AI was negotiating a raise of about $500M at a $3B pre-money valuation, with Premji Invest expected to lead. We saw that report relayed by TheNextWeb and StartupFortune rather than in Bloomberg’s original. Bloomberg’s funding scoops are generally reliable, and this one mostly checks out against what was later announced.
The arithmetic lines up: $3B pre-money plus $500M in new money gives the $3.5B post-money figure Forbes reports. The one difference is the lead. Bloomberg’s report pointed to Premji alone. The announced round, per TechCrunch and Forbes, has Premji leading alongside Khosla and Eclipse. Treat the Bloomberg report as useful context for how the deal came together, not as a separate source of facts. Everything that matters now comes from the announcement.
What we don’t know yet
Most of the questions an engineer would ask are still open.
How many robots are actually doing work. “Built 30 to 40” and “piloting with a handful of customers” are company statements. We have no independent count of deployed units, no customer names, and no description of the tasks. A pilot can mean a robot doing real shifts on a line. It can also mean a cell in a corner running supervised trials. The reporting so far doesn’t tell us which.
Whether glove-based training holds up outside demos. The company’s pitch is that capturing human hand motion with gloves gives robots human-level dexterity. That approach could be a cheap, scalable source of manipulation data. It could also run into the usual problems: the gap between a human hand and a robot hand, and the gap between collected data and messy real environments. We found no independent evaluation, published success rates, or third-party test results that settle this.
Whether the GENE-26.5 demo was autonomous. We could not establish whether the May 2026 demo with Wuji Tech hands ran autonomously or under teleoperation. That distinction matters a lot in dexterous manipulation. Teleoperated demos show what the hardware can do. Autonomous demos show what the model can do. Until Genesis AI says which it was, the demo is evidence of hand hardware, not of learned dexterity.
Timing. Reports differ. One says Eno mass production is aimed at 2027. Another says targeted deployments begin by the end of 2026. No official source confirms either.
What it means
Investors are pricing Genesis AI on its approach, not its fleet. A $3.5B valuation against 30 to 40 robots in undisclosed pilots is a bet that glove-captured human dexterity can become a scalable data pipeline for manipulation. Manipulation is one of the hardest unsolved problems in robotics, and it is where general-purpose robots will either earn their keep or not.
The bet isn’t crazy. Data is the bottleneck in robot learning, and anything that collects useful manipulation data cheaper than fleets of teleoperated robots is worth pursuing. Leads like Khosla and Eclipse, plus Nvidia on the investor list, signal that serious money finds the thesis credible.
But funding confirms a funding round, not a capability. The things that would show this approach works are not public yet: autonomous task success rates, named customers willing to talk, uptime numbers, and a clear answer on how the GENE-26.5 demo was run.
Credits and sources
By RWT Robotics Editorial Team
- Published









