1X Technologies reveals Neo robot hand — human-level grip strength, shipping this year

1X Technologies is revealing the new hand for its Neo humanoid robot, matching average human grip strength at roughly three times the force of competing robot hands on the market. Neo ships in 2026 — Bernt Øyvind Børnich won't commit to a specific date, but says the company will keep its promise.

“The hand is roughly the same strength as an average human... If you look at the general hands in the market right now, this thing is roughly three times as high force as the other hands... We have promised that we are going to ship this year, and we will ship this year.”

The hand

The hand is built entirely in-house at 1X's California factory, where the company designs and produces its own motors, tendons, sensors, and electronics. The design philosophy is biomimetic: 1X engineered the fingers to replicate the nonlinear compliance of a human hand, not for aesthetics but because it closes the gap between robot and human interaction with the world. The idea is that if the hand interacts with objects exactly as a human hand does, the company can train on existing internet video rather than relying purely on bespoke robot-specific data.

Grip strength is a direct constraint on which tasks a robot can learn. Børnich's argument is that if a hand is a third as strong as a human's, it simply cannot access a third of human experience — and model intelligence is a direct function of experiential diversity. The hand can hold a barbell loaded to 150 pounds, the same weight Neo can deadlift.

Platform and distribution

1X plans to open Neo as a platform, inviting third-party developers to build on a single homogeneous hardware base. Børnich says that kind of shared benchmark infrastructure doesn't exist in robotics today and is worth preserving, though he leaves the door open to hardware partnerships if manufacturing scale is the outcome.

Ramp expectations

Børnich is direct that physical AI will diffuse more slowly than software — no Waymo-style rollout goes global overnight. His bet is that the ramp is slower early but the eventual ceiling is far higher than software, because robots building more robots leads to full automation of the physical substrate: data centers, chip fabs, energy infrastructure, mining. He puts meaningful scale two to three years out, though he acknowledges it could be a decade. The early demand signals he cites are wet labs wanting robots to run scientific experiments autonomously, hospitality, elderly care, and the home.