1X Technologies announced a 25 Degree of Freedom (DOF) tendon-driven hand for its NEO humanoid platform, resolving the physical limitations that historically restricted robotic manipulation. This system transitions humanoid hands from rigid end effectors to read-write instruments that perceive environments through physical touch. Global businesses planning humanoid deployments in 2026 must evaluate how these engineering changes alter software requirements, as physical capability no longer limits AI model outputs. See our Full Guide on humanoid design evolution.

How does the 1X Tendon Drive achieve force transparency in robot hands?

The 1X Tendon Drive achieves force transparency by operating quasi-direct-drive tendons at low gear ratios of 5:1 to 15:1. Typical robotic hands use gear ratios of 100:1 or 200:1, which introduce immense friction that blocks physical contact forces from reaching the motor. The NEO hand bypasses this mechanical bottleneck, ensuring that any force applied to a finger flows back through the transmission to be measured by the motor itself.

The mechanics of low-friction transmissions

Traditional robot joints are write-only devices where controllers command a position, but receive no feedback when encountering obstacles. The low gear ratio of the 1X Tendon Drive reduces friction to a level where the transmission is fully backdrivable. When a finger presses against a surface, the external force reverses the drive mechanism, allowing the motor to detect the exact resistance. This bi-directional flow of physical force turns every joint into an active force sensor without requiring heavy external sensor arrays.

Proprioception through closed-loop joints

Every joint in the NEO hand operates on a closed-loop system that tracks position and torque simultaneously. The hand measures the difference between target positions and actual physical resistance. This continuous feedback loop allows the robot to build a spatial and tactile map of its immediate environment. The hand reads the physical world with high dynamic range, distinguishing between a delicate pinch and a firm grasp.

Why are 25 degrees of freedom necessary for humanoid robot hands?

A 25 degree-of-freedom hand is necessary to allow humanoid robots to execute complex, multi-point manipulation tasks without relying on simplified geometric paths. By integrating 22 fully actuated degrees of freedom in the fingers and palm, along with three degrees of freedom in the wrist, the NEO hand replicates the kinematic range of a human hand. This mechanical design expands the operational vocabulary of the robot far beyond simple pick-and-place routines.

Expanding the developer command vocabulary

Simple two-finger grippers limit developers to three basic physical commands: picking, placing, and pushing. These movements are executed blindly, as the machine cannot sense orientation or slip. The 25 degrees of freedom on the NEO hand allow developers to write applications for complex physical interactions, such as threading a needle or turning a key. Software models can now leverage the high mechanical dexterity to execute continuous adjustments during a task.

Active palm and wrist integration

The inclusion of three degrees of freedom in the wrist and active joints in the palm allows the NEO hand to alter its grasp shape dynamically. This capability prevents fragile objects from breaking while ensuring a secure hold on heavy or slippery items.

How does the NEO hand integrate tactile feedback with physical robustness?

The NEO hand integrates tactile feedback with physical robustness by combining high-bandwidth shear-sensing skin with reflex-driven control loops. This combination allows the robot to detect slip and adjust its grip tension within milliseconds, protecting both the hand and the handled objects from damage. The physical design uses durable materials designed to withstand millions of contact cycles during industrial operations.

Tactile and shear-sensing skin technology

The fingers of the NEO hand are wrapped in a tactile skin that measures pressure, pressure changes, and shear forces across the entire surface. This sensor configuration detects the exact moment an object begins to slide out of a grip. Instead of relying solely on visual confirmation from cameras, the robot processes tactile data streams to determine if an object is slipping or tilting.

Reflex-driven control loop speeds

To prevent drops, the hand utilizes localized reflex loops that operate independently of the main AI perception stack. If the shear sensors detect slip, the joint controllers instantly increase gripping force before the object can fall. This high-bandwidth loop ensures that the robot responds to physical disturbances faster than central vision processing allows.

Physical durability for continuous operations

Repetitive physical contact causes rapid wear in traditional robotic components. The NEO hand is engineered for high durability, allowing it to perform thousands of probing cycles daily. By distributing forces through the tendon-drive system and using robust synthetic skins, 1X ensures that the hand can interact with rough surfaces without degrading its sensory accuracy.

Key Takeaways

  • B2B developers can bypass the limitations of two-finger grippers, using 25 degrees of freedom to program complex tactile tasks.
  • Force transparency via low gear ratios (5:1 to 15:1) allows the NEO hand to operate as a bi-directional sensor, reading physical resistance in real time.
  • High-bandwidth shear-sensing skins and local reflex loops prevent object slippage faster than standard vision-based perception stacks.