“The test of an innovation is always what it does for the user” — Peter F. Drucker*

A data ecosystem for scalable AI-based robotics
Physical AI becomes commercially relevant when robots can combine perception, simulation, planning and action with trusted access to the right industrial data. RoX—the Digital Ecosystem for AI-Based Robotics—is turning this proposition into practice.

What is RoX?
RoX is a €52 million pre-competitive R&D initiative running from 2024 to 2027, supported by Germany’s Federal Ministry for Economic Affairs and led by ABB. The consortium brings together more than 20 partners to develop shared foundations, standards and reusable building blocks that would be difficult to create individually. It includes leading robotics providers such as ABB, Dürr, Siemens, Schunk and Rheinmetall; AI research experts and service providers including the German Research Center for Artificial Intelligence (DFKI), the German Aerospace Center (DLR), Fraunhofer, Google’s Intrinsic and T-Systems; and pioneering industrial users including Mercedes-Benz and Würth.

Why RoX matters for industry and the C-suite
The challenge is not robot capability alone. Adoption remains constrained by costly, complex integration, programming and reconfiguration, with systems integration often exceeding 50% of deployment costs—especially in dynamic production environments and for smaller companies. AI can improve flexibility and productivity, but depends on specialised, proprietary data distributed across machines, processes, companies and systems. RoX addresses both barriers by combining AI-powered applications with governed access to reusable data and software components. For business leaders, this creates three strategic opportunities:
     •  Better economics: Reduce bespoke integration and reconfiguration through reusable, interoperable components
     •  Ecosystem scale without loss of control: Share data and capabilities across companies while retaining sovereignty over access and usage
     •  A path to agentic automation: Enable AI agents to discover data and services, coordinate tasks and connect with business processes under trusted governance

How RoX works: Layered architecture and customer-driven protyping
RoX follows a market-oriented, results-driven development strategy. Building on Peter Drucker’s customer-centric logic, the consortium starts with practical use cases defined by end users and works backward to determine what is required across the software stack and within the physical robotic cell. Rather than developing another isolated robotics solution, RoX builds a three-layer data ecosystem:
     •  Dataspace network layer for trusted and governed data exchange
     •  Reusable data products, semantic models and federated catalogues
     •  AI-powered robotics apps such as Pick & Pack
T-Systems leads the Digital Robotics Ecosystem work package and provides the operational Dataspace-as-a-Service infrastructure based on open-source Eclipse Tractus-X. Separate development and stable operating environments allow the consortium to experiment continuously without disrupting its live demonstrators.

Results and milestones
RoX ties demonstrator delivery to major industry events. These fixed milestones focus collaboration across partners, require infrastructure, data products, applications and robotic cells to work together, and provide market feedback for the next development cycle.
•  Hannover Messe — April 2026: Second-generation demos of Pick & Pack, Quality, and Teaching consuming ecosystem data through federated catalogue capabilities. Intrinsic, a Google company, demonstrated a Pick & Pack simulation application using live data assets to help close the sim-to-real gap. T-Systems report | LinkedIn
•  Japan outreach — March–April 2026: RoX and its AI-dataspace approach were presented through webinars hosted by Kyoto University KUPAC and the Robot Revolution & Industrial IoT Initiative. T-Systems report | RRI event
•  Automatica — June 2025: Three live robotic cells demonstrated Pick & Pack, Quality Inspection and Teach & Assemble, complemented by a first ecosystem application visualising cross-company data flows. T-Systems report | LinkedIn
•  Hannover Messe — April 2025: The first major public checkpoint connected RoX with the broader Manufacturing-X and cloud-edge agenda. T-Systems report | LinkedIn
•  Project launch — 2024: RoX was launched in Berlin by Federal Minister Robert Habeck, followed by the consortium kickoff at Dürr.

Publications
•  RoX overview — 2-page introduction
•  Data Spaces and AI — 1-page RoX callout in the IDSA publication
•. RODEOS – Robotic Data Ecosystem Semantic Model: Bridging siloed robot systems

* Drucker, P. 1985. Innovation and entrepreneurship: Practice and principles. Harper & Row: p. 252