Genesis AI Releases GENE-26.5 Robotic Hand

The robot then applies electrical tape to wires, solves a Rubik’s cube, grabs a straw from a container and separates a plastic cup from a stack, all in fully autonomous mode at 1x speed. The company, based in San Carlos and Paris, emerged from a year of quiet development.

It builds the entire technology stack: a robotics-native foundation model, a 1:1 human-like robotic hand, a noninvasive data collection glove that captures motion, force and touch from human workers, and a simulator that compresses weeks of physical experiments into minutes.

Genesis AI raised $105 million from Eclipse, Khosla Ventures, Bpifrance, Eric Schmidt and Xavier Niel.

The company’s approach centers on capturing human dexterity directly at scale through the glove to overcome the data scarcity that has limited robotics foundation models. CEO Zhou Xian told TechCrunch the firm decided to develop the full stack only after determining it needed control over the hardware.

The demonstration has drawn significant attention on X. Genesis AI emphasizes the video runs at normal speed without acceleration, unlike many robot videos. If the system performs equally outside controlled environments, the company believes it represents a substantial advance.

A few weeks before the launch, Tyler Habowski and Yonatan Robbins of Kyber Labs described a different reality in a podcast interview. Habowski, who previously worked at SpaceX, said there are literally zero robot hands deployed right now doing routine work.

He added that the best hands cost hundreds of thousands of dollars and break frequently. Kyber Labs began with a hardware-first strategy after its founders evaluated existing options. The company developed torque-transparent actuation that uses motors themselves as force sensors, eliminating the need for separate tactile or force-torque sensors.

This design aims to replicate how humans rely on force feedback rather than precise positioning.

AI bets that the foundation model and diverse dexterity dataset will become the competitive moat, with hardware serving mainly as a delivery mechanism. The company’s hand was built because suitable options were unavailable commercially. Its long-term goal is general-purpose physical AI.

Kyber Labs, by contrast, focused first on durable, force-sensitive hardware suitable for narrow but practical deployments. Its initial commercial system operates in a clinical laboratory where it handles pipettes, uncaps tubes and mixes samples on a stationary arm without legs or wheels.

The firm describes its path as pragmatic rather than aiming for immediate general-purpose autonomy. >"There are literally zero robot hands deployed right now doing routine work. " — Tyler Habowski, Kyber Labs co-founder (Forbes) Both companies agree that durability at reasonable cost will determine commercial success.

A robotic hand must cost less than the human worker it replaces and survive continuous operation. Questions remain about manufacturing cost, service life under load and maintenance requirements for the Genesis system, which the company has not yet disclosed.

Industry participants emphasize the importance of moving beyond laboratory demonstrations. Yonatan Robbins of Kyber Labs noted that a hand must withstand repeated tasks such as washing after cracking eggs to avoid spreading residue. Tyler Habowski compared current robotics development to SpaceX’s iterative launch process versus NASA’s traditional engineering approach.

He argued the field needs widespread deployment to generate the real-world data required to identify and correct shortcomings. Both the Genesis breakfast-preparation demo and Kyber’s clinical lab system represent early steps that must be tested at scale.

The broader robotics sector continues to show rapid progress in controlled settings. Yet the transition to millions of units performing everyday tasks such as laundry folding, food preparation and mechanical assembly still depends on hardware reliability, software generalization and economics that compete with human labor.