Siemens built an AI agent that designs humanoid robot grippers from a text prompt

Siemens built an AI agent that designs humanoid robot grippers from a text prompt

John KoetsierNovember 21, 2026
ShareXLinkedInFacebook
Siemens and Chinese humanoid maker AGIBOT have demonstrated an AI workflow that takes a text description of a robot task and returns a validated, print-ready gripper design in roughly 1 day, down from weeks of manual CAD work.

From text to a robot gripper

Type what a robot needs to grip and what it needs to do with it. Get back a gripper, designed, simulated, validated, and ready to print.

Magic?

Nope, just a new agentic AI workflow that Siemens unveiled with Chinese humanoid maker AGIBOT on November 16, ahead of Formnext 2026 in Frankfurt. An engineer enters a natural-language description of the target object and the task. The AI agent selects the gripper type, adapts the geometry inside Siemens' Designcenter software, runs simulation and virtual validation, checks the robot interface, and generates print-ready files for additive manufacturing.

Siemens says the full design cycle compresses from weeks to roughly 1 day.

How the pipeline works

The technical enabler is Model Context Protocol (MCP) integration, which allows the AI agent to reach into and orchestrate functions across Siemens' engineering software stack without manual hand-offs between tools. The same digital thread that carries the original task description also carries the validated design into production, with no re-entry of data at any stage.

Karsten Heuser, VP of Additive Manufacturing at Siemens: "Instead of manually designing variants in a time-consuming process, agentic workflows can orchestrate engineering tasks based on specific requirements."

5 steps, 1 prompt, 1 validated output file.


The Humanoid Technology Expo is where the builders, integrators, and enterprise buyers shaping this space come together. Early registration is open.


Why grippers are a bottleneck

Humanoid robots don't use 1 gripper. They use many.

Every new application, every new object, every new environment can require a different end effector: different geometry, different compliance, different mounting interface. Traditional engineering cycles for each variant take weeks. At the production volumes humanoid makers are now targeting, that is a real constraint.

(This is one reason why, by the way, great hands are so valuable: they will work on anything.)

AGIBOT, based in Shanghai, had produced 10,000 humanoid robot units by March 2026. Its robots work in factories and warehouses, environments that change frequently. The company's gripper requirements change with them.

AGIBOT's side

AGIBOT is one of the most active humanoid makers in China, targeting factory and logistics applications with a fleet that requires application-specific end effectors. The Siemens workflow is meant to give AGIBOT engineers the ability to respond to new task requirements without burning weeks per gripper variant.

What Siemens gets: a live customer pushing the limits of an emerging AI design stack, and the credibility of a partnership with one of the fastest-moving humanoid manufacturers in the region.

The Erlangen production cell

Beyond the design workflow, Siemens is building an adaptive additive manufacturing production cell at its Gerätewerk Erlangen electronics factory in Germany. Partners are EOS, supplying the printing equipment, and AM Solutions, handling automated post-processing. The intent is that the same digital thread feeding the AI design workflow feeds directly into this cell for physical production.

"The next stage in Industry 4.0 is Industrial AI: the shift from connected and automated factories to intelligent, adaptive manufacturing," says Marie Niehaus-Langer, CEO of EOS.

David Soldan of AM Solutions adds a caveat that matters most for real-world production:

"Only when downstream processing steps are seamlessly integrated into the production chain can additive components be manufactured efficiently and reproduced at scale."

What comes next

Siemens is also enhancing its AM I Navigator tool with industrial AI, giving manufacturers a structured method to find and scale AI applications across their own additive operations. The Erlangen cell is the proof of concept. The broader claim is that any manufacturer adopting the Siemens stack can run the same workflow.

How widely AGIBOT deploys the system in production, and whether Siemens signs additional humanoid partners, the company has not said. Formnext runs through November 22. The booth is Hall 12.1, D119.

Sources

Share this article

ShareXLinkedInFacebook

In this story

Follow to get an alert when they're in the news

Siemens
Company · Supplier
Germany
AgiBot
Company · Humanoid maker
China
Newsletter

Humanoid (almost) Daily

The humanoid robot news that matters, with context from HumanoidMegaHub's tracking. Free, almost every day.

Subscribe on LinkedIn

Keep reading

Written by

John Koetsier
John Koetsier

I tell stories with data and build startups. I've raised angel and VC funds, led software companies and startups, and chronicled the rise of the mobile economy. Currently I write for Forbes, host several podcasts, and am starting up HumanoidMegaHub. I was TUNE's mobile economist and Singular's VP of Insights, built the research team at VentureBeat, and managed teams creating software for partners like Intel and Disney. In past lives I built websites, e-learning platforms, and mobile apps, and, a long time ago, was named to Folio's top 100 of the media industry's "most innovative entrepreneurs and market shaker-uppers.”