Subscribe

Siemens and AGIBOT show an AI agent that designs robot grippers from a text prompt

Siemens and AGIBOT are showing a workflow in which an AI agent designs a 3D-printed gripper for a humanoid robot from a written description of the object and the task. The agent picks a gripper type, adapts the geometry, simulates it and prepares it for printing. It is a demo at the Formnext trade fair, and no performance figures were given.

Image: via 3D Printing Industry

Siemens and the Chinese humanoid maker AGIBOT are showing a workflow in which an AI agent designs a gripper for a humanoid robot from a text prompt, 3D Printing Industry reports. The user describes the object to be handled and the task the robot has to do. The agent then picks a suitable gripper type, adapts its geometry in Siemens' Designcenter software, simulates it, and checks the robot interface and the handling. After that it prepares the data for 3D printing. Siemens is exhibiting the workflow at the Formnext 2026 trade fair in Frankfurt.

The demo covers the design side only. For production, Siemens says it is working with EOS and AM Solutions on an adaptive production cell at its electronics factory in Erlangen, which is described as planned. Siemens says the approach could open complex engineering work to more users, because they no longer have to operate each tool themselves. The report gives no figures on how well the generated grippers perform or how long the process takes.

Why it matters

A humanoid's hands decide which jobs it can do, and custom grippers are slow to design by hand. If an agent can turn a task description into a printable, simulated gripper, a robot could be fitted for a new job much faster. For now this is a trade fair demo from the companies themselves, so the claim still has to be shown on a real factory floor.

ELiA Weekly

The robot news that mattered, every Sunday.

The videos, the stories explained in plain words and the headlines worth your time. Free, in your inbox.

More news

All news