AddiWire

Additive manufacturing news and European prices, with independent expert commentary

Software

Siemens demos AI agent that designs printable robot grippers from a written task

Published 1 sourceSummary written with AIDeutsche Version

Siemens and AGIBOT showed an AI agent that turns a written task description into a gripper design for a humanoid robot, checks it virtually and prepares the data for additive manufacturing. The workflow is part of a wider Digital Thread for AM, shown at Formnext 2026 in Frankfurt. Siemens is also planning an adaptive production cell with EOS and AM Solutions in Erlangen to carry such designs into physical production.

Key facts
  • Demo use case: a gripper for an AGIBOT humanoid robot
  • Designcenter and MCP let the agent run engineering steps
  • Adaptive production cell planned with EOS and AM Solutions in Erlangen
  • Shown at Formnext 2026, Hall 12.1, Booth D119

How does the agent design a gripper?

The user describes what the robot has to pick up and what it must do with it. From that input the AI agent selects a suitable gripper type and adapts its geometry in Siemens' Designcenter. It then simulates the result and checks the robot interface and the handling of the object. Through MCP, the agent can call functions of the engineering software and coordinate the individual steps. At the end, the data is prepared for additive manufacturing, so the gripper can be validated virtually before anything is printed.

Siemens says this could make complex engineering workflows accessible to more people, since they would not need to operate each tool and step themselves.

What is planned on the production side?

Siemens is working with EOS and AM Solutions on an adaptive production cell at its Erlangen electronics plant. The cell is meant to combine additive manufacturing, automation and software into one process chain that can handle varying components and still run repeatably and at scale. Post-processing was included in the cell concept from the start, as downstream steps decide whether printed parts can be made efficiently in volume.

Siemens names quality, traceability, security and human responsibility as requirements for industrial AI use. It has also added Industrial AI to its AM I Navigator, a tool for finding and ranking AI use cases in AM design, engineering and production, from which companies can derive pilot projects and a roadmap.

How does it compare with other AI tools for AM?

Synera uses AI agents to operate existing AM software, automating file repair, orientation, support generation and slicing, including Materialise Magics through its SDK. Engineers at the BMW Additive Manufacturing Campus used Synera for lightweight robot grippers, built from automated workflows and topology optimization. Ai Build's Aibuild 2.0 accepts text commands that generate toolpaths. These tools take text input for different purposes: Siemens' agent starts from the task and selects and adapts the design, while a Talk to AI command can set a parameter such as layer height.

What it means

For small and mid sized manufacturers this is a direction rather than a product to buy. The gripper is a demonstration, and the production cell in Erlangen is still planned, so no cost per part or qualification route can be derived yet. The practical value lies in the idea of validating a design virtually before printing, which could shorten iteration for custom tooling and end-of-arm parts. Skills in structured requirements and data management will matter more than tool-by-tool know-how. This continues Siemens' Formnext push for an AI-driven AM workflow reported earlier.

Earlier on AddiWire

Frequently asked questions

What does the Siemens AI agent do?

It takes a written description of the object and task, selects a gripper type, adapts the geometry in Designcenter, simulates it, checks the robot interface and handling, and prepares the data for additive manufacturing.

Who is building the adaptive production cell?

Siemens is working with EOS and AM Solutions on a cell at its electronics factory in Erlangen. It is planned to combine AM, automation and software, with post-processing included from the start.

Where can the workflow be seen?

Siemens is showing it at Formnext 2026 in Frankfurt, Hall 12.1, Booth D119.

Read the original reports
  1. 3D Printing Industry, 20 Nov 2026