📊 Full opportunity report: EuroHPC. The compute substrate. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

EuroHPC’s compute infrastructure currently supports European AI projects, including several regional AI Factories and flagship supercomputers, but faces structural limitations for frontier-scale model training. The €20 billion InvestAI Facility aims to expand capabilities through AI Gigafactories, with procurement decisions ongoing, marking a critical phase in Europe’s AI infrastructure development.

EuroHPC’s current compute substrate underpins numerous European AI initiatives, with systems like JUPITER, LUMI, and Leonardo ranked among the world’s top supercomputers. These systems enable mid-sized model training, exemplified by Apertus on Alps, which demonstrates operational feasibility at the 70-billion parameter scale. However, the existing infrastructure is insufficient for training frontier-class models, which require the scale targeted by the upcoming AI Gigafactories. The EuroHPC framework comprises 19 AI Factories across 21 countries and 13 AI Factory Antennas, forming regional ecosystems that provide access to supercomputing resources, data services, and talent support. These facilities are primarily concentrated in wealthier member states, such as Germany, Italy, Spain, and France, creating geographic and economic disparities. The €20 billion InvestAI Facility is designed to establish up to five AI Gigafactories capable of training trillion-parameter models, addressing the current capability gap. The selection process for these Gigafactories is ongoing, with decisions expected through summer 2026, aligning with the EU AI Act enforcement window in August. This timing underscores the strategic importance of the compute substrate’s readiness to support Europe’s AI ambitions.

Implications of EuroHPC Infrastructure for Europe’s AI Leadership

The current EuroHPC compute substrate confirms Europe’s ability to support mid-sized AI models, but highlights significant structural challenges for scaling to frontier AI training. The ongoing development of AI Gigafactories aims to bridge this gap, shaping Europe’s competitive position in AI research and deployment. The geographic concentration of flagship systems may also influence the equitable distribution of AI capabilities across member states, impacting broader European AI policy goals.

EuroHPC’s Evolving Infrastructure and Strategic Framework

Since its creation in 2018, EuroHPC JU has coordinated Europe’s supercomputing efforts with a €10 billion investment (2021-2027), including the deployment of top-ranked supercomputers like JUPITER, LUMI, and Leonardo. The institutional architecture features regional AI Factories, national gateways, and plans for large-scale AI Gigafactories supported by the €20 billion InvestAI Facility. These initiatives are aligned with Europe’s broader strategy to develop sovereign AI capabilities, as articulated in recent policy updates and the expansion of the EuroHPC mandate through Council Regulation (EU) 2026/150.

The infrastructure supports numerous projects, including the seven-essay European sovereign-LLM framework, which relies on the compute substrate for operational deployment. However, despite operational success at the mid-sized model level, structural limitations remain for frontier AI training, which the new Gigafactory framework aims to address.

“The EuroHPC infrastructure is operationally credible at the AI Factory tier for mid-sized model training but structurally insufficient for frontier-class training, which the €20 billion AI Gigafactory framework aims to address.”

— Thorsten Meyer

Unresolved Challenges and Future Infrastructure Developments

It remains unclear how quickly the AI Gigafactory procurement process will conclude and how effectively these new facilities will address the structural limitations. Additionally, the impact of geographic concentration and heterogeneity in hardware and software remains to be fully assessed as deployment progresses.

Next Steps in Europe’s AI Infrastructure Expansion

Procurement decisions for the AI Gigafactories are expected to finalize through summer 2026, with operational deployment likely in the subsequent years. These developments will determine whether Europe’s compute substrate can support frontier AI training at scale and how regional disparities will evolve. The EU AI Act enforcement in August 2026 will further influence strategic priorities and investments.

Key Questions

What is the current capacity of EuroHPC’s supercomputers for AI training?

EuroHPC’s flagship systems like JUPITER, LUMI, and Leonardo support mid-sized model training, with Apertus demonstrating operational capability at 70 billion parameters. However, they are not yet sufficient for training trillion-parameter models.

What are the main limitations of the current EuroHPC infrastructure?

The infrastructure is structurally insufficient for frontier-scale AI training, faces hardware heterogeneity and software complexity issues, and is geographically concentrated in wealthier member states, creating disparities across Europe.

How will the AI Gigafactory framework address current limitations?

The €20 billion InvestAI Facility aims to establish up to five AI Gigafactories capable of training trillion-parameter models, addressing the capability gap and scaling Europe’s AI training capacity.

When are the AI Gigafactory decisions expected?

Procurement and selection processes are ongoing, with key decisions expected through summer 2026, aligning with policy and regulatory timelines.

What impact might geographic concentration have on Europe’s AI development?

The current concentration of flagship supercomputers in wealthier countries may deepen regional disparities, potentially influencing the equitable distribution of AI capabilities across Europe.

Source: ThorstenMeyerAI.com