On 20 February, during Officine d’Intelligenza – 1st National Forum on Artificial Intelligence for Industry, the Italian Institute of Artificial Intelligence (AI4I) publicly presented, through its Director Antonio Calegari, its first multi-year Strategic Plan for 2026–2030.
The document sets out a five-year framework for direction, programming and performance measurement, aligned with three-year and annual operational cycles. More than a planning exercise, the Strategic Plan establishes the structural link between mission, organisation, measurable objectives and financial sustainability. Strategy, governance and execution are explicitly connected: general objectives are translated into operational indicators, and indicators into concrete activities.
In accordance with the Statute, the Plan was presented by the Director to the Supervisory Committee, which holds institutional oversight over strategic choices and resource allocation. Following a positive opinion from the Scientific Committee, the Plan was formally approved on 31 October 2025.
A Dual Mission: Industrial Adoption and National Technological Capacity
The Strategic Plan orients AI4I’s activity along two complementary trajectories:
- Accelerating AI adoption across industrial sectors
- Strengthening national capability to develop and produce AI technologies
The first trajectory concerns operational diffusion: artificial intelligence must become part of companies’ ordinary processes. The second concerns the consolidation of proprietary expertise, reusable components and technological autonomy.
The Plan identifies a structural shift in the technological cycle: value is increasingly generated not only in model design, but in industrialisation, integration into production processes, lifecycle management, security, and governance of data and AI systems. Reliability, resilience and operational continuity become central.
An Integrated, Multi-Divisional Architecture
The Strategic Plan defines AI4I as an integrated system covering the full value chain — from knowledge generation to industrial deployment.
As illustrated in the integrated architecture presented in the slides, the Institute is organised into six operational units:
- Research, Development & Engineering (RDE) Center
- AI4I Foundry
- System for User Knowledge (SUK)
- Institute for Advanced Study (IAS)
- AI4I Academy
- AI4I StartGarden
Together, these units form a continuous chain in which each component prepares and enables the next: research feeds engineering, engineering supports deployment, deployment scales through infrastructure, and adoption is reinforced through training, ecosystem building and entrepreneurial valorisation.
Research, Development & Engineering: From Prototypes to Deployment
The RDE Center combines specialised laboratories, engineering units and a Deployment function.
Its structure, detailed in the slides, includes domain-oriented laboratories in:
- Software, computation and cybersecurity
- Product design and optimisation
- Automation and autonomous systems
Among the laboratories already established are:
- Advanced Materials & Engineering Design (AMED)
- Physical Holistic Intelligence (PHI)
- AI Security (AIS)
Each laboratory follows an agile archetype, combining scientific depth with adaptability to technological and market evolution.
The RDE Scientific Advisory Board includes internationally recognised scholars and industry leaders, reinforcing the Institute’s scientific credibility and global positioning.
The growth trajectory is equally ambitious. According to the projections, the number of laboratories is expected to rise progressively to approximately 30 by 2028–2030, with laboratories accounting for roughly 80% of operating costs at steady state.
Revenue targets for RDE activities indicate a progressive increase in external income through commissioned research, joint labs, competitive grants and AI Factory services, reaching 62% coverage of R&D costs by 2030.
AI4I Foundry: A Dedicated HPC Platform for AI
At the infrastructural core of the Institute stands the AI4I Foundry, a proprietary AI-optimised HPC platform.
As specified in the technical slide, the Foundry includes:
- 80 NVIDIA Blackwell B200 GPUs (AI-optimised partition)
- 68 NVIDIA Hopper H200 GPUs (general-purpose partition)
- 5 Petabytes of storage
With a total capital investment of approximately €10 million, the Foundry ranks among the largest AI-dedicated machines available in Italy.
The platform is structured to serve internal research, projects with companies, SUK producers and external enterprises and is complemented by managed services covering:
- HPC compute-as-a-service
- AI model training and inference
- GPU testing and validation sandbox
- MLOps and CI/CD integration
- Secure sovereign cloud infrastructure
- Pilot-to-production scaling support
The Foundry is scheduled to enter into operation in the first half of March, with a target utilisation rate of approximately 80% within two years.
SUK: A Platform for Matching Demand and Supply
The System for User Knowledge (SUK) functions as AI4I’s marketplace and matchmaking infrastructure.
Its four core functions include:
- Collection of industrial needs
- Mapping of available AI solutions
- Selection of integrators
- Project accompaniment
At the beginning of 2026 SUK had reached a first critical mass of:
- 92 producers
- 124 AI solutions
The growth targets are significant: over 8,000 engaged companies, more than 600 producers and a rhythm of one match per day by 2030.
IAS, Academy and StartGarden: Knowledge, Skills and Entrepreneurship
The Institute for Advanced Study (IAS), developed in partnership with Fondazione Compagnia di San Paolo, positions AI4I within the international scientific debate.
The IAS hosts seminars, round tables, lectio magistralis and intensive schools with internationally recognised scholars such as Marco Pavone, Bill Dally, Pietro Perona and others contributing to a structured programme that combines scientific excellence with industrial dialogue.
From 2026 onward, IAS activities are expected to include:
- 1–2 lectio magistralis per year
- 12 seminars/roundtables
- 2–3 intensive programmes (winter/summer schools)
The Academy and StartGarden complete the chain by fostering managerial and technical skills for AI adoption and supporting the creation and growth of startups and spin-offs.
Roadmap, Sustainability and European Dimension
The Strategic Plan translates its mission into measurable milestones.
Key elements include:
- Progressive expansion of R&D laboratories to approximately 30 units
- Growth of external revenues beyond €10 million in the medium term
- Deployment and scaling of the Foundry
- Integration into the European AI Factory trajectory and positioning toward future AI Gigafactories
Financial projections indicate stable public endowment complemented by increasing external income and reinvestment in computational capacity.
Partnerships play a central role. The model includes commissioned research, joint labs, competitive grants and membership agreements with industrial and institutional partners contributing to research orientation, deployment capacity and ecosystem building.
From Strategy to Execution
Presented publicly at Officine d’Intelligenza, the 2026–2030 Strategic Plan marks a transition from institutional set-up to full operational scaling.
It formalises AI4I as:
- A research and engineering centre
- An infrastructural platform
- A marketplace for AI solutions
- An international scientific hub
- A training and entrepreneurial ecosystem
Above all, it defines a measurable path to bridge research, engineering and industrial adoption — positioning AI4I as a national infrastructure designed to strengthen competitiveness and technological capacity in artificial intelligence over the coming decade.
