Significant public investment is being put into the next phase of Italy’s high-performance AI infrastructure, with a new AI supercomputer being commissioned at the IT4LIA AI factory in Bologna.
The effort is being led by the European High Performance Computing Joint Venture (EuroHPC JU) in partnership with E4 Computer Engineering and Dell Technologies.
The system forms part of a €290 million program covering procurement and long-term operation. The funding will be split equally between support from the Digital Europe Program (DEP) and co-financing from the Italian Ministry of University and Research.
Once deployed, the platform, which will deliver peak AI inference performance of over 160 exaflops, will anchor Italy’s AI factory within the wider European network, providing massive AI computing power and expanding access to startups, small and medium-sized businesses, and researchers working on data-intensive applications.
Anders Jensen, Executive Director of EuroHPC JU, commented:
“With the IT4LIA system, we are making significant investments in cutting-edge infrastructure that enables startups, small businesses, and the research community to develop and deploy advanced AI applications at scale.
“IT4LIA will not only expand computational capacity, but also strengthen collaboration across the European innovation ecosystem and play a key role in strengthening Europe’s competitiveness and leadership in AI.”
What is IT4LIA AI Factory designed to achieve?
The IT4LIA AI Factory is part of the broader EuroHPC initiative to establish a distributed network of AI factories across Europe. A total of 19 facilities have been deployed, supported by 13 additional antenna sites to expand service regionally.
Its core function is practical: making high-performance AI computing accessible to companies and researchers who lack the resources to build or access such systems.
The Factory model focuses on applied use cases, especially for small and medium-sized businesses and startups, and provides both infrastructure and technical support.
IT4LIA has already been operational since April 2025 using existing systems, including Leonardo supercomputers.
The new AI supercomputer will transition the site into a full-scale AI factory with a gradual change in functionality and dedicated infrastructure for training and inference workloads.
How is an AI supercomputer built and what are its characteristics?
This system is integrated by E4 Computer Engineering and manufactured by Dell Technologies.
It is based on a water-cooled architecture using NVIDIA GB200 NVL4 technology, combined with NVIDIA Grace CPUs and Blackwell GPUs, interconnected via Quantum X800 InfiniBand networking.
This configuration is designed for large-scale AI workloads, prioritizing throughput, scalability, and energy efficiency.
Performance, expected to exceed 160 exaflops, is specifically focused on AI inference rather than general purpose computation, aligning the system with real-world deployment needs.
A unified data layer is provided through the VAST data platform, enabling streamlined data access and management across workloads.
The system also includes a dedicated inference partition that integrates Axelera AI’s accelerators and SiPearl’s CPU, strengthening the contribution of European technology in the stack.
Who will benefit from this AI factory and how?
The primary users will be startups and small businesses, which will be able to access advanced AI computing without the capital burden of building their own infrastructure.
However, access is not limited to industry. The broader research community will also be able to use the system to develop, test, and deploy AI applications.
This factory provides both vertical and horizontal services. Sector-specific support includes agritech, cybersecurity, climate and weather, and manufacturing.
These are complemented by cross-cutting tools such as secure data management, compliance validation, and metadata services.
Training and skills development will also be a focus. The AI Factory is structured not just as an infrastructure, but as an ecosystem that supports adoption from initial experimentation to deployment.
Why is this important for Europe’s AI strategy?
The IT4LIA project is positioned as part of a broader effort to build a secure and sovereign AI infrastructure within Europe.
The system aims to ensure that AI developments are aligned with local regulatory frameworks by incorporating compliance with EU data protection and privacy standards.
The dedicated European partition within the AI supercomputer is intended to optimize inference workloads while maintaining control over sensitive data. This reflects a strategic focus on autonomy at both the hardware and software layers.
In practical terms, the AI Factory model lowers barriers to innovation while keeping important functions under European control.
The implementation in Bologna marks a transition from fragmented resources to a coordinated, high-capacity AI infrastructure designed for sustainable industrial and research applications.
Source link
