Skip to main content
Publications

HPC7, supercomputer at the service of energy

HPC7 is Eni’s latest high-performance computing system and, together with HPC6, enables the company to surpass the Exascale threshold. The supercomputer supports energy transition activities including CO₂ storage studies, next-generation battery development, biofuel optimisation and fusion research. Located within Eni’s Green Data Center, HPC7 combines advanced computing performance with energy-efficient liquid-cooling technology and is also being opened to companies, startups and research centres.

  • HPC7 and HPC6 together enable Eni to surpass the Exascale computing threshold.
  • HPC7 supports energy transition applications including CO₂ storage studies, next-generation batteries, biofuels and fusion research.
  • The system combines CPUs, GPUs and liquid-cooling technology to maximise computational performance and energy efficiency.
  • HPC7 strengthens Eni’s technological leadership in industrial high-performance computing.
  • Eni is making its supercomputing infrastructure and AI capabilities available to companies, startups and research centres

Our supercomputing system is among the most powerful in the world and it accelerates our transformation.

Data center with server racks illuminated by blue lights.

The HPC7 High Performance Computing (HPC) system, completed and commissioned in June 2026, significantly enhances the computing power of Eni’s Green Data Center in Ferrera Erbognone, in the province of Pavia. Combined with the previous HPC6 system, HPC7 enables us to surpass the Exascale threshold, equivalent to a computing power of more than one quintillion (one billion billion) mathematical operations per second (1048 PFlop/s sustained and 1467 PFlop/s peak). 

With a sustained computing capacity of more than 571 and a peak performance of over 861, HPC7 is the world’s leading technological achievement among industrial supercomputers, ranking sixth globally and second in Europe in the TOP500 ranking. The new infrastructure is complemented by the previous HPC6 system, inaugurated in November 2024, which retains a sustained performance (Rmax) of 477 PFlop/s and a peak computing capacity of more than 606 PFlop/s. 

The entire system continues to rely on liquid-cooling technology, optimising heat dissipation and reducing overall energy consumption, in line with the most advanced efficiency standards in supercomputing. 

  • 1 of 1

     

    Supercomputing opens up to the innovation ecosystem

    Eni is making its supercomputing infrastructure, AI services and specialist expertise available to companies, start-ups and research centres. The initiative accelerates time to market, the development of advanced models and digital solutions, while strengthening industrial competitiveness.
    Go to press realease

A strategic asset supporting business evolution

Main features

Agribusiness and agricultural supply chains
Agribusiness and agricultural supply chains
Carbon management
Carbon management
Exploration and production
Exploration and production
Renewable
Renewable
Scientific research
Scientific research

Industrialisation

Artificial intelligence

Carbon management

Fluid dynamics

Geophysical exploration

Molecular modelling

Cineca
CNR Lecce

How does it work?

With a peak computing power of 861 PFlop/s¹, the new HPC7 system is based on a hybrid architecture combining CPUs and GPUs, featuring almost 3,400 computing nodes² and nearly 14,000 GPUs to maximise computational performance and energy efficiency.

It is used to optimise industrial operations, improve the accuracy of geological and fluid dynamics studies for CO₂ storage, develop next-generation batteries, optimise the biofuel value chain and simulate plasma behaviour in magnetic confinement fusion.

Its high computing power accelerates Eni’s transformation, supporting the development of new high-potential businesses linked to the energy transition. It also strengthens synergies between Eni and its Satellite companies and creates strategic value for new external collaborations.

  • (1) PFlops: a measure of computing power equivalent to 10¹⁵ floating-point operations per second.
  • (2) Node Composition: each node consists of four AMD Instinct™ MI300A Accelerated Processing Units (APUs), each combining 24 AMD EPYC™ Zen 4 CPU cores and 228 CDNA3 GPU compute units, for a total of 96 CPU cores and 912 compute units per node. Each node is equipped with 512 GB of HBM3 memory.

Features and performance

The HPC7 is a computing cluster, i.e. a set of computers working together to multiply overall performance.
  • 861
    mln of billions

    mathematical operations performed per second

  • 3,480

    total CPUs

  • 13,920

    total GPU cards

Illustrated timeline of supercomputing history, comparing major global milestones and Eni developments from 1948 to 2026, highlighting the evolution of computing power and artificial intelligence.
Illustrated timeline of supercomputing history, comparing major global milestones and Eni developments from 1948 to 2026, highlighting the evolution of computing power and artificial intelligence.

Discover more

Worker on board the ship looks at the horizon
VISION

Natural gas: reliable and diversified supplies

Woman and businessman talk in the office
PRODUCTS

By supplying energy to companies, we support their sustainable growth

Download document Download document