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

EuroHPC’s compute infrastructure currently supports regional AI Factories and mid-sized model training, but it is not yet sufficient for frontier-class AI training at scale, according to recent analyses. This confirms that the existing European supercomputing platform is operationally capable but faces structural limitations as Europe prepares for the upcoming AI Gigafactory selection process.

The EuroHPC Joint Undertaking (JU) has established a network of 19 AI Factories across 21 European countries, supported by a €10 billion investment in supercomputing infrastructure from 2021-2027. These Factories, along with 13 AI Factory Antennas, form the core of Europe’s regional AI ecosystem. The Compute Concentration Audit: When Sovereign Wealth Funds Notice Three Companies Own the Frontier.

Recent deployments include systems like Leonardo (ranked #10 globally in TOP500) and JUPITER (#4 worldwide), which demonstrate Europe’s operational capacity for mid-sized AI model training, exemplified by Apertus training a 70-billion-parameter model on Alps. However, the infrastructure’s heterogeneity—due to multi-vendor hardware and software fragmentation—and concentration in wealthier member states highlight structural issues.

The €20 billion InvestAI Facility aims to create up to five AI Gigafactories capable of training trillion-parameter models, addressing the capacity gap identified in prior analyses. The selection process for these Gigafactories is ongoing, with decisions expected by summer 2026, aligning with the EU AI Act enforcement window of August 2, 2026.

Implications of EuroHPC Infrastructure for Europe’s AI Strategy

The current EuroHPC compute substrate confirms operational readiness for mid-sized AI model training, supporting Europe’s regional AI initiatives and smaller-scale projects. However, it reveals critical structural limitations for frontier AI training, which the €20 billion InvestAI framework aims to address through the development of dedicated AI Gigafactories. These developments are pivotal for Europe’s competitiveness in AI, especially as procurement decisions and policy enforcement deadlines approach in summer 2026.

EuroHPC’s Infrastructure and Europe’s AI Ambitions

Since its creation in 2018, the EuroHPC Joint Undertaking has coordinated Europe’s supercomputing efforts, pooling resources from the EU and member states. The 2021-2027 investment plan allocates €10 billion for infrastructure and AI Factories, with 19 Factories and 13 antennas established across the continent. These systems have already supported projects like Minerva, Apertus, and Aleph Alpha, which rely on EuroHPC’s compute substrate for training models ranging from hundreds of millions to 70 billion parameters.

Despite operational successes, structural issues—such as hardware heterogeneity, software complexity, and geographic concentration—limit the infrastructure’s capacity for frontier AI training. The ongoing development of the AI Gigafactory framework, with a €20 billion fund, is designed to address these limitations by creating large-scale, dedicated facilities for trillion-parameter models.

“The EuroHPC infrastructure confirms operational capability for mid-sized models but exposes structural gaps for frontier-class training, which the Gigafactory framework aims to fill.”

— Thorsten Meyer

Unresolved Challenges in Europe’s Compute Infrastructure

It remains unclear how quickly the AI Gigafactory framework will be implemented and whether the selected facilities will fully resolve the capacity and heterogeneity issues. The precise timeline for procurement decisions and the operational readiness of new Gigafactories are still developing, with final outcomes expected later in 2026.

Next Steps in EuroHPC and AI Gigafactory Deployment

The European Commission will finalize the selection of AI Gigafactory sites through 2026, with decisions expected by summer. These facilities are intended to address the current capacity gaps and support the EU’s AI policy objectives, including compliance with the upcoming AI Act enforcement starting August 2, 2026. Monitoring procurement progress and infrastructure deployment will be critical in assessing Europe’s AI competitiveness.

Key Questions

What is the current capacity of EuroHPC systems for AI training?

EuroHPC systems like Leonardo and JUPITER support mid-sized models up to around 70 billion parameters, demonstrating operational capacity for regional AI projects.

Why are the existing EuroHPC systems insufficient for frontier AI models?

The infrastructure faces structural limitations such as hardware heterogeneity, software complexity, and geographic concentration, which hinder scaling to trillion-parameter models required for frontier AI training.

What is the role of the €20 billion InvestAI Facility?

It aims to fund up to five AI Gigafactories capable of training trillion-parameter models, addressing the capacity gap for Europe’s AI ambitions.

When will the AI Gigafactory sites be selected?

The selection process is ongoing, with decisions expected by summer 2026, ahead of the EU AI Act enforcement deadline in August 2026.

How might the current infrastructure impact Europe’s AI leadership?

While operational for mid-sized models, current limitations could slow Europe’s progress in frontier AI development unless the Gigafactory framework is successfully implemented and deployed.

Source: ThorstenMeyerAI.com