Genesis AI’s Revolutionary Leap into Robotics with GENE-26.5
Genesis AI, a startup that raised a $105 million seed round to build foundational AI for robotics, has revealed its first model, GENE-26.5, and it comes with surprise hands. In a demo video, the company demonstrated various advanced tasks performed by a set of robotic hands it designed in-house.
The Full-Stack Approach to AI and Robotics
“The model has always been the goal, because a better model means better intelligence,” Zhou Xian, co-founder and CEO of Genesis, told TechCrunch. But the company quickly realized it needed to control the hardware. “So we decided to go full stack,” he said.
Other well-funded companies operate at the intersection of AI and robotics, such as Physical Intelligence and Skild AI. Zhou also acknowledged that “there are probably 50 or 100 robotic hand companies.” But he and his co-founder Théophile Gervet hope that building their own business will give them the upper hand.
The Human-Like Robotic Hand
The main difference is that Genesis’ hand is the same size and shape as a human hand – rather than the two-fingered grippers used by many robotics companies – narrowing the gap with real-world conditions.
“This allows us to collect much more data than before, to train a model capable of performing many more tasks,” said Gervet, former research scientist at Mistral AI and now president of Genesis.
From Cooking to Complex Tasks
Of all the physical manipulation tasks shown in the video below, Gervet’s favorite is cooking, because it proves that the robot was able to complete a long series of difficult tasks, like cracking an egg or slicing a tomato. But Genesis also tasked its robots with making smoothies, playing the piano, and solving the Rubik’s Cube – a robotic gadget.
Commercial Applications and Data Collection
Other tasks, like lab work, are closer to what commercial applications of Genesis’ technology could be. But what’s happening behind the scenes is just as important: The startup has also developed a sensor-laden glove that functions as a real-life duplicate of its robotic hand, collecting more usable data.
“Our idea was that if we could design a robotic hand that attempts to mimic a human hand as closely as possible, we could instantly unlock huge amounts of human data without having to worry about what people call the ’embodiment gap’ in robotics research,” Zhou said.
Innovative Data Strategies
Others have tried their hand at this problem; the main novelty is how Genesis combines this with its model. The current version is called GENE-26.5 for May 2026, but Zhou expects many iterations, thanks to the simulation he developed. “The real bottleneck for model iteration speed is evaluation, so this helps us significantly speed up model training,” he said.
Beyond simulation, however, data will be key to training models that can help robots perform more tasks. This is also where the Genesis Gauntlet could come in handy. Gervet said that, unlike bulky data collection devices that get in the way, they are just as lightweight and easy to wear as the safety gloves already used in many industries, while being relatively inexpensive to manufacture.
Industry Implications and Future Outlook
“We are currently in talks with many customers, and a lot of the benefit of a glove would be that for the first time you can wear the data collection device when you are doing your daily job, whether that is a lab technician for the pharmaceutical industry or for manufacturing,” Gervet said. This would also be supplemented by “egocentric video data” – people filming themselves completing a task.
It remains to be seen, however, whether workers would be happy to wear the gloves and cameras that could train the robots to replace them, and whether they will receive additional compensation for this training. This will play out between Genesis customers and their employees, Gervet suggested. “We haven’t worked out the details yet,” he said.
Either way, they might decide not to share that data with the startup, the founders acknowledged. But the startup also has its own means to build its “human skills library”: it could also pay third-party partners to collect data. Its model is already trained on “massive quantities of human-made Internet videos,” according to a press release that does not mention compensation.
Combined with its simulation system, this could help Genesis reduce the costs of its technology for real-world applications like the one it demonstrated. “This marks an important milestone for their team and for the robotics industry in general,” said former Google CEO Eric Schmidt, who invested in the startup.
Strong Funding and Expansion
In July 2025, just a few months after its creation, the startup emerged from hiding with a $105 million funding round co-led by Eclipse and Khosla Ventures, with additional backers including Bpifrance, HSG, and individuals like Schmidt, but also Xavier Niel, Daniela Rus, and Vladlen Koltun.
This funding allowed Genesis to increase its workforce. With offices in Paris and California, it has also expanded to London. “One of the main reasons we decided to expand into Europe is that there is a huge density of talent across the continent,” Gervet said. Its team of 60 people is split between “40-45% in Europe and 50-55% in the United States,” and the startup is currently recruiting across all three locations.
Future Developments
Besides hiring, the company also plans to soon unveil its first general-purpose robot, which, Zhou told TechCrunch, will be a complete robot, not just hands. But he insisted the roadmap is still the same.
“Our goal is to build the most capable robotic system,” he said.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
“`
