📊 Full opportunity report: EuroHPC. The compute substrate. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
EuroHPC’s current compute infrastructure supports mid-sized AI models but faces structural limitations for frontier-scale training. The €20 billion AI Gigafactory plan aims to address these gaps, with ongoing procurement and policy developments shaping Europe’s AI future.
EuroHPC’s existing supercomputing infrastructure is operationally capable of supporting mid-sized AI model training but is not yet sufficient for frontier-scale models, according to recent analysis. The €20 billion AI Gigafactory framework, designed to close this gap, is in the procurement and selection phase, with decisions expected by summer 2026, making this a critical period for Europe’s AI ambitions.
The EuroHPC Joint Undertaking (JU) has established a network of 19 AI Factories and flagship supercomputers such as JUPITER, LUMI, and Leonardo, which currently support AI research at a scale up to approximately 70 billion parameters, exemplified by Apertus on the Alps system. These facilities enable regional ecosystems for AI development, startup support, and data services, forming the backbone of Europe’s AI infrastructure.
However, the infrastructure’s capacity for training large, frontier-class models—those exceeding hundreds of billions of parameters—remains unproven at scale. The €20 billion InvestAI Facility aims to establish up to five AI Gigafactories capable of trillion-parameter training, addressing this capability gap. The selection process for these Gigafactories is ongoing, with a timeline set for summer 2026, coinciding with the EU AI Act enforcement window.
Structural issues have emerged, including hardware heterogeneity—CUDA, ROCm, multi-generation hardware fragmentation—that increases software complexity and operational overhead for European AI developers. Additionally, the concentration of flagship systems in wealthier member states like Germany, Italy, Spain, and France creates geographical and economic disparities, potentially reinforcing structural inequalities within the EU’s AI ecosystem.
EuroHPC.
The compute
substrate.
€10 billion AI Factories + €20 billion AI Gigafactories. 19 AI Factories + 13 Antennas. JUPITER #4, LUMI #9, Leonardo #10. Federation Platform shipped April 15. The compute substrate underlying every project in the seven-essay framework — and the three structural complications the framework didn’t address directly.
This is the eighth standalone essay in the European sovereign-LLM track and the first Tier 2 expansion piece. The prior seven essays documented six institutional answers plus the integrative synthesis framework. Every one of those projects depends operationally on the EuroHPC compute substrate or a national-equivalent. Apertus trained on Alps (10,752 GH200 superchips, 4,096 GPUs). OpenEuroLLM allocated millions of GPU hours across multiple EuroHPC systems. Minerva trained on Leonardo. AMÁLIA on Deucalion. Mistral on commercial cloud + ASML strategic-investor partnership. Aleph Alpha historically on alpha ONE + now Schwarz Group STACKIT + €11B Berlin DC. The compute substrate is the unifying infrastructure question the seven-essay framework didn’t address directly. Summer 2026 is the operational moment when the substrate’s strategic positioning is determined.
Two tiers. One scale gap.
The EU policy framework operates two structurally distinct programmatic tiers. The bifurcation explicitly acknowledges that current AI Factory tier infrastructure is insufficient for frontier-class model training. The AI Gigafactory framework is the EU policy framework’s operational response to the structural capability gap Finding 1 from the synthesis essay surfaces empirically.
Two Channel SXM2 Expansion Board Builts for Data Center GPUs Featuring Advanced 300G Cooling Solution Servers GPU Accelerators Board
Engineered for, the SXM2 two GPU expansion baseboard 300G supports two SXM2 GPUs ( V100) with integrated NVLink…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Six flagships. Six chromatic cross-references.
The flagship EuroHPC systems crystallize the substrate underlying the seven-essay framework. Three rank in the global TOP500 top 10. Two are exascale (one operational, one deploying 2026). All six are project-cross-referenced in the seven-essay framework. The chromatic register of each system maps to its project cross-reference.
30B+ trained
LUMI users
training
Factory
2026
70B
Handbook of Research on Methodologies and Applications of Supercomputing (Advances in Systems Analysis, Software Engineering, and High Performance Computing)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Three cohorts. 21 European countries.
The AI Factory selection has expanded rapidly through December 2024 – October 2025 across three cohorts. 13 AI Factory Antennas in 7 EU Member States plus 6 partner countries complete the framework. The Antennas are the institutional infrastructure connecting Apertus (Switzerland) and other partner-country projects to the EuroHPC framework.
CEREBRAS WSE-3: LARGE-SCALE AI TRAINING ON WAFER-SCALE ARCHITECTURE: Build Trillion-Parameter LLMs with Massive On-Chip Memory, Simplified Programming, and Cluster-Scale Performance
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Three complications. Three policy gaps.
The compute substrate analysis surfaces three structurally distinct complications. These are not criticisms of EuroHPC — they are the operational realities the strategic discourse should integrate. The Federation Platform partially addresses the first; the AI Factory Antennas framework partially addresses the second; the AI Gigafactory framework explicitly addresses the third.
European supercomputing infrastructure
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Summer 2026. Three deadlines simultaneously.
The June 2026 AI Gigafactory selection process, the August 2 EU AI Act enforcement window, and the Q4 2026 EuroHPC Federation Platform second release all converge in summer 2026. This is the operational moment when the European sovereign-AI compute substrate’s strategic positioning is determined for the 2027-2029 horizon.
4 weeks ago
from now
moment
from now
from now
months
from now
The work is real across the EuroHPC framework. Substantial infrastructure built. 19 AI Factories operational or in deployment. 13 Antennas connecting smaller member states. EuroHPC Federation Platform shipped April 15, 2026. Apertus 70B operationally demonstrates Alps-tier training. The structural complications are also real. Heterogeneity hidden cost. Geographical concentration. Scale-tier bifurcation. Both can be true at once. Summer 2026 is the operational moment when the European sovereign-AI compute substrate’s strategic positioning is determined.
Implications of Infrastructure Limitations for Europe’s AI Leadership
The current EuroHPC compute substrate confirms that Europe can support mid-sized AI models but faces significant structural hurdles for large-scale, frontier AI training. Addressing these limitations through the AI Gigafactory initiative is essential for Europe’s competitiveness in advanced AI development. The infrastructure challenges—hardware heterogeneity and geographical concentration—may influence policy decisions and resource allocation, impacting Europe’s ability to lead in frontier AI innovation and democratize access across member states.
EuroHPC’s Infrastructure and Strategic Developments in European AI
Since its creation in 2018, EuroHPC JU has coordinated Europe’s supercomputing efforts, with a €10 billion investment planned for 2021-2027. The network includes regional AI Factories, national gateways, and flagship supercomputers like JUPITER (ranked #4 globally), LUMI (#9), and Leonardo (#10). Recent expansions under Council Regulation (EU) 2026/150 have broadened the JU’s mandate to include AI Gigafactories and quantum technologies. The current infrastructure supports models up to approximately 70 billion parameters, but the need for larger, trillion-parameter models has driven the €20 billion InvestAI Facility and the ongoing Gigafactory selection process, with decisions expected by mid-2026.
“The EuroHPC infrastructure framework is operationally credible at the AI Factory tier for mid-sized model training but structurally insufficient for frontier-class training, which is what the €20 billion AI Gigafactory framework aims to address.”
— Thorsten Meyer
Unresolved Challenges in Hardware Diversity and Geographical Equity
It remains unclear how quickly the Gigafactory procurement process will progress, the specific locations of the selected facilities, and how effectively the infrastructure will scale for trillion-parameter models. Additionally, the impact of hardware heterogeneity and regional concentration on operational efficiency and equitable access is still being evaluated.
Upcoming Milestones in AI Infrastructure Deployment and Policy
Key next steps include the finalization of the AI Gigafactory selection process in summer 2026, followed by the deployment of new facilities. The enforcement of the EU AI Act in August 2026 will also influence operational priorities. Monitoring how these developments address current structural limitations will be critical for assessing Europe’s AI leadership trajectory.
Key Questions
What is the current capacity of Europe’s supercomputing infrastructure for AI?
Europe’s supercomputing infrastructure supports models up to around 70 billion parameters, as demonstrated by systems like Apertus on Alps, but is not yet capable of training trillion-parameter models at scale.
What is the purpose of the AI Gigafactory framework?
The €20 billion InvestAI Facility aims to create up to five large-scale AI facilities capable of training trillion-parameter models, addressing the current capability gap for frontier AI development in Europe.
What structural challenges does Europe’s AI infrastructure face?
Major challenges include hardware heterogeneity (CUDA, ROCm, multi-generation hardware), which increases software complexity, and geographical concentration of flagship supercomputers in wealthier member states, which may exacerbate regional disparities.
When will the final decisions on AI Gigafactory locations be made?
The selection process is ongoing, with expected decisions and announcements scheduled for summer 2026.
How does this infrastructure impact Europe’s AI policy goals?
It is central to Europe’s strategy to become a leader in frontier AI, but current limitations mean that significant infrastructure expansion and policy adjustments are necessary to meet ambitious goals.
Source: ThorstenMeyerAI.com
