Siemens links industrial AI to additive manufacturing

Siemens links industrial AI to additive manufacturing

Siemens links industrial AI to additive manufacturing of robotic grippers. Its demonstrated workflow moves from task descriptions through design checks to production preparation, while a separate adaptive manufacturing cell is still being developed.


Siemens has demonstrated an industrial AI workflow that converts a robotic handling task into a gripper design prepared for additive manufacturing. The sequence combines geometry selection and adaptation, simulation, robot-interface checks and preparation of manufacturing data. A separate project at the company’s Electronics Factory in Erlangen aims to connect additive production with automated finishing. The gripper workflow has been demonstrated; the proposed adaptive manufacturing cell is still being developed.

Working with robotics company AGIBOT, Siemens uses the example of a gripper for a humanoid robot. An engineer describes the object to be picked up and the required handling task. An AI agent then selects a suitable gripper type, modifies its geometry in Siemens’ Designcenter software, simulates the arrangement and checks the robot interface and handling operation. The digital design is subsequently prepared for additive manufacture. The company plans to demonstrate this sequence at Formnext 2026 in Frankfurt.

The agent accesses engineering software functions through an integration using Model Context Protocol, allowing it to coordinate operations across tools rather than produce a drawing or text recommendation alone. Each step still relies on correct task requirements and the validity of the underlying design and simulation models. Siemens has not published physical endurance test results showing that the demonstrated gripper satisfies every application requirement, so virtual validation and suitability for deployment should not be conflated.

Gripper design changes with the geometry, mass and surface characteristics of the item being handled. Contact points, clearance and the forces applied to a component determine whether a design is appropriate for its task, while the robot interface places limits on how it can be mounted and moved. Repeating CAD edits, simulation and manufacturing preparation for each new part variant can create work across several engineering systems. Siemens’ demonstration coordinates these stages but does not supply comparative figures for engineering time saved or error rates.

Additive manufacturing changes the production constraints because a new geometry can be built from a revised digital file without the dedicated tooling associated with many conventional fabrication processes. The suitability of a printed gripper still depends on material choice, build orientation, mechanical performance and the finishing needed at functional surfaces. A component that passes a virtual handling check may require physical inspection and qualification before deployment in a production environment.

At its Erlangen Electronics Factory, Siemens is working with EOS and AM Solutions on an adaptive production cell. The planned environment is intended to join additive manufacturing equipment, automation, downstream processing and software into a coordinated production chain. It is designed to accommodate different component designs without rebuilding each process around a separate manual setup. The company has not provided a commissioning date, cycle-time results or evidence that the installation has achieved sustained commercial throughput.

Post-processing is an important part of the proposed arrangement. Printed components often require powder or support removal, surface treatment, cleaning or inspection before they can be used, depending on the material and process. Automating only the printing stage can leave these tasks as a bottleneck when production switches between component designs. AM Solutions is contributing expertise in downstream processing, while EOS provides additive manufacturing knowledge and equipment capability.

As designs change, the corresponding production information must also remain consistent. A modified CAD model can alter build preparation, material allocation, processing settings or inspection criteria; passing an out-of-date file between those stages would defeat the intended automated workflow. Siemens identifies traceability, quality, security and human responsibility as requirements for its approach. Those controls will need to be demonstrated in operation rather than assumed from the presence of an AI agent or a connected data system.

Siemens has previously demonstrated factory AI applications concerned with industrial simulation and production processes. The AGIBOT project instead starts with the requirements of an individual manufactured component. Its sequence ends with manufacturing preparation, leaving the physical build and downstream verification to other equipment and processes. The Erlangen cell is intended to connect those subsequent stages, but the two initiatives are not evidence of an already deployed autonomous design-to-production system.

An updated version of Siemens’ AM I Navigator incorporates industrial AI to help identify suitable use cases, prioritise pilots and plan wider deployment of additive manufacturing applications. The tool’s assessment role is separate from physical qualification of the gripper or commercial validation of the Erlangen cell. No quantified return on investment has been published for either initiative, and the manufacturer has not set out a general availability timetable for customers adopting the complete sequence.

At Formnext, Siemens will show how an agent moves from the handling requirement to a design ready for printing. The Erlangen project will have to demonstrate repeatable operation when different component geometries require different manufacturing and finishing steps. Part quality, traceable changes and the time spent between design approval and usable output remain the practical measures by which such a production chain can be evaluated.


Stories for you


  • Exosens secures order for 25,000 intensifier tubes

    Exosens secures order for 25,000 intensifier tubes

    Exosens will supply 25,000 image intensifier tubes to Theon Sensors. The firm 4G+ order supports European night vision manufacturing, with options covering another 35,000 units and deliveries expected to begin late this year.


  • Epiroc completes acquisition of French mining supplier Aramine

    Epiroc completes acquisition of French mining supplier Aramine

    Epiroc has completed its acquisition of French mining supplier Aramine. The transaction brings an established underground equipment parts and remanufacturing business into its service organisation, but excludes the separately established ARAMAC business.