Since their introduction in the European Commission’s AI Innovation Package (July 2024), 19 AI Factories have been launched across 16 EU countries (with 5 more member states involved) and 13 antennas, to cover all EU territory. These AI Factories act as ecosystems combining computing power, data, and talent to build large AI models and foster AI driven innovation. [AG1]

BDVA, as member of the EuroHPC JU representing Data and AI and industrial users, has been contributing from the very beginning to the successful establishing of the AI Factories and their connection to data ecosystems, which will be facilitated by the upcoming Data Labs.

This session will explore this interplay, trying to clarify roles, prevent overlaps, and define how AI Factories, Data Labs and Data Spaces can be integrated into a coherent ecosystem aligned with EU strategic goals.

Objectives

More specifically, this session aims to:

• Provide policy background and expectations on data labs
• Review the data capabilities that the different AI factories are planning to offer, and their ongoing collaborations with data initiatives
• Start building bridges between AI Factories and data spaces in several sectors, exploring potential collaboration models
• Identify technical and non-technical challenges in the short term, including how to prepare data specifically for AI (AI-ready data products), data governance of the overall ecosystem (licenses, quality, ….) and others.

Agenda

Part 1: 9:00 – 10:30 CET

• Objectives of the session and BDVA’s role

Introduction to Data Labs

• Presentations from AI Factories
• Presentation HPC Federation

Part 2: 10:50 – 11:50 CET

• Intro and recap from previous part
• Presentation from data spaces
• Panel discussion with AI Factories, Data spaces and HPC federation

Session organised by Jeanette Nilsson (RISE), Ana García Robles (BDVA), and Daniel Alonso Román (BDVA)