Michel Iñigo Ulloa from the AASDS-IND project on the importance of interoperability, data-sharing and open source
Voices of 8ra
This interview series highlights the leaders shaping the Multi-Provider Cloud-Edge Continuum across Europe. We explore the strategies and challenges driving innovation within the 8ra Initiative – set against the backdrop of shifting political priorities, rapid technological change, and evolving societal and economic needs that are redefining Europe’s digital future.
November 2025
The 8ra Team in Conversation with Michel Iñigo Ulloa, Senior Innovation & Technology Manager, MONDRAGON Corporation
European factories generate vast amounts of data, but currently systems rarely align. Within the 8ra Initiative, MONDRAGON Corporation, a Spanish cooperative business group in the areas of industry, finance, retail and knowledge, is working to make industrial AI collaborative by design. In this interview Michel Iñigo Ulloa, Senior Innovation & Technology Manager at MONDRAGON and project lead in the AASDS-IND project, describes how the team connects diverse equipment across plants, applies federated learning so models train where data is created, and places latency-critical decisions at the edge – all while keeping ownership and governance with each company.
The approach prioritises interoperability and purpose-bound data sharing over centralisation. Early trials focus on quality control and predictive maintenance – use cases where latency matter and confidentiality is non-negotiable.
Michel, what is MONDRAGON’s role within the 8ra Initiative?
“We lead the project AASDS-IND in 8ra’s Workstream 4 “Advanced Applications”. We’re assembling a data space reference architecture made of three parts: (1) an Asset Administration Shell (AAS) “digital twin” kit to represent industrial equipment consistently and standardized; (2) an interoperability kit – middleware that connects machines and IoT platforms from multiple vendors; and (3) a federated-learning kit so companies can improve models and productivity without moving raw data. We’re testing this on quality control and predictive maintenance.
In the virt8ra testbed, we integrate with the broader stack: industrial edge (e.g., Siemens), orchestration (e.g., OpenNebula) and compute providers such as Arsys – and we contribute infrastructure in the Basque Country.”
You emphasise interoperability as the key challenge. Why is that so central?
“Getting heterogeneous systems to work together is a main challenge. Each manufacturing plant mixes technologies and suppliers. Our interoperability kit provides a single middleware layer, based on Asset Administration Shell, so assets ‘speak’ a common language. Without that layer, AI and data sharing don’t scale beyond a single line or factory.”
How do data spaces help beyond interoperability?
“In a data space, each company sets enforceable rules: who can use what, for which purpose, under which conditions. We put that sovereignty into the architecture, so collaboration doesn’t require surrendering control.”
What does this have to do with Europe’s digital sovereignty?
“Digital sovereignty is something Europe must achieve if it wants to act independently in the digital domain. Within the 8ra Initiative, we contribute to this goal in several ways.
We contribute by building on open-source projects developed and maintained in Europe, keeping companies’ ownership of data and know-how and validating solutions in an open, federated Multi-Provider Cloud-Edge Continuum.
It’s about creating the capability to innovate and collaborate globally, while maintaining control and trust locally.”
What does federated learning change in this context?
“Federated learning allows us to train models where the data is created. Only model parameters or insights travel – not raw data. In industrial use cases, that’s very relevant.
For predictive maintenance, multiple factories with similar assets contribute to more robust models. For quality control, sites align performance across similar lines.
And because critical industrial environments operate in real time – often within milliseconds – the learning process must be efficient and distributed. You can’t afford latency or long feedback loops. That’s why placing computation at the edge is so important.”
You’ve mentioned that in industrial operations, real time truly means milliseconds. How does the Multi-Provider Cloud-Edge Continuum support such fast decision-making?
“Machine signals should be collected continuously and sometimes need to react within milliseconds. That’s why inference and certain control loops at the edge should be placed close to the equipment, and use the cloud for aggregation and retraining. The Multi-Provider Cloud-Edge Continuum is about the right workload in the right place – and about being able to reconfigure quickly.“
How does this work tie into collaboration across 8ra?
“We rely on capabilities from other workstreams – management and orchestration, security, resilience. In virt8ra, we validate end-to-end: device → edge → data space → model → orchestration – with partners across providers and technologies.”
Why commit to open source?
“Transparency and independence. We use – and contribute to – projects like Eclipse Dataspace Connector (EDC) and Catena-X/Tractus-X where they fit industrial needs. If you depend on a component, you should help maintain it. That’s how we keep the technology sustainable and aligned with European needs.”
A shared foundation for Europe’s industrial cloud-edge future
MONDRAGON’s work in AASDS-IND shows how the 8ra Initiative translates European digital-sovereignty goals into practical solutions and demonstrates how interoperability and governance can come together in real industrial environments.
As Michel Iñigo Ulloa put it:
“8ra gives Europe the chance to improve industrial operations – open, resilient and sovereign.”
It’s a reminder that Europe’s competitive strength lies not in scale alone, but in its ability to build technology that reflects its own values – trust, collaboration and innovation across a shared Cloud-Edge Continuum.
About Michel Iñigo Ulloa
Michel Iñigo Ulloa is Senior Innovation & Technology Manager at MONDRAGON Corporation. He leads AASDS-IND within 8ra’s Workstream 4 “Advanced Applications”, focusing on Asset Administration Shells, data spaces and federated AI for quality control and predictive maintenance.