LUMI supercomputer

LUMI is EuroHPC JU’s flagship supercomputer and the 11th fastest globally (Top500 list published in June 2026). LUMI is also one of the most eco-efficient data centers in the world. You can read more about LUMI’s sustainability on the Sustainable future section.

LUMI, an HPE Cray EX system, has a sustained computing power of 380 petaflops (HPL, High-Performance Linpack) in its final configuration. LUMI is also one of the world’s leading platforms for artificial intelligence.

LUMI facts

		LUMI’s sustained computing power (HPL) is 380 petaflops (sustained performance) in its final configuration.
		LUMI’s computing power is equivalent to the combined performance of 1.5 million of the latest laptop computers. These would form over a 23-kilometer-high tower.
		LUMI is using 100% renewable energy. LUMI’s waste heat is used to heat hundreds of households in the city of Kajaani.

Used technologies: The supercomputer achieves its high performance with a large number of nodes with accelerators (GPUs). In addition, the system is complemented by a CPU-only partition, IaaS cloud services, and a large object storage solution.

Read also a blog post about the system architecture: May we introduce LUMI and the news about the system architecture.

A description of the LUMI system architecture and the different hardware partitions on LUMI docs.

LUMI-AI supercomputer

LUMI-AI will gradually replace the current LUMI supercomputer. LUMI-AI will enable breakthroughs in AI training and inference, as well as large-scale scientific simulations. The system will be delivered by Bull and it will be located in CSC’s new data center in Kajaani, Finland. It is expected to deliver a tenfold increase in artificial intelligence (AI) capacity and nearly double the high-performance computing (HPC) capability of the current LUMI supercomputer. The system is scheduled for deployment in the second half of 2027.

LUMI-AI system architecture

The liquid-cooled supercomputer will be based on Bull’s latest BullSequana XH3500 architecture, specifically designed for advanced AI workloads. It will be built on next-generation AMD Instinct™ MI430X GPUs alongside state-of-the-art 6thGen AMD EPYC™ 256-core processors, providing flexibility across diverse computing workloads.

The other technology partners are IBM, providing an ambitious storage solution based on their Storage Scale technology, as well as Nokia, providing data center networking expertise to build a network that complements the BXI interconnect by Bull.

The system will be complemented by additional compute and storage resources, maximising its overall performance and usability. A key differentiator is its multi-tenant architecture, designed to meet the needs of diverse AI communities, combined with extensive API-based access to support modern development environments.

Here you will find design phase images of the system’s architecture, software-defined clusters and storage solution. See also the media gallery for further images.

Read also press release: Bull selected to deliver LUMI-AI supercomputer, powering next-generation AI workloads and beyond