National Academic Infrastructure for Super­computing in Sweden

What National Academic Infrastructure for Super­computing in Sweden does

NAISS (National Academic Infrastructure for Supercomputing in Sweden) is the government-mandated national e-infrastructure providing high-performance computing, storage, and cloud resources to Swedish academic researchers. It is financed by the Swedish Research Council (Vetenskapsrådet) with co-funding from the Knut and Alice Wallenberg Foundation (KAW) for AI/ML resources, and operates as a non-commercial entity serving universities and research institutes across Sweden. Resources are provided free of charge to eligible PIs (≥60% appointment at a Swedish HEI recognized by VR) through merit-based allocation rounds administered by the National Allocations Committee (NAC) via the SUPR portal.

The infrastructure portfolio comprises multiple supercomputer systems: the new Arrhenius (382 GPU nodes with 4x NVIDIA GH200 each, 66.8 PFLOPS, #42 on the June 2026 TOP500 list, plus CPU, SENS, Disk, Flash, and PCD partitions); Berzelius (NVIDIA DGX SuperPOD with A100 and H200 nodes at NSC, KAW-funded); Dardel (HPE Cray EX with GraceHopper and MI250X partitions at PDC/KTH); Alvis (dedicated AI/ML GPU cluster at C3SE/Chalmers); Bianca (sensitive-data cluster at UPPMAX); Tetralith (NSC/Linköping); plus Kebnekaise, Pelle, COSMOS, and others across six regional centers. Storage spans Swestore/dCache, Klemming (Lustre, 12 PB), Berzelius Storage (VAST Data), Cephyr (Ceph), and Mimer (all-flash + Ceph for AI). The Swedish Science Cloud provides OpenStack IaaS across three geographic regions.

Go-to-market is direct-to-researcher with no pricing: SUPR handles project registration, allocation applications, and resource provisioning. Distribution is mediated by the six regional HPC centers (NSC, PDC, UPPMAX, C3SE, HPC2N, LUNARC), which provide local support and training. Outreach includes an annual User Forum (hybrid; 2026 in Stockholm with NVIDIA's Jeff Hammond as keynote), a quarterly newsletter edited by Director Erik Lindahl, and free training/workshops including CodeRefinery. Strategic positioning is anchored by the 2026 Swedish LLM partnership with WASP and KAW on Berzelius, the Mimer AI Factory under EuroHPC JU, and decommissioning of legacy systems (Tetralith/Alvis 1 July 2026, Dardel 1 January 2027) in favor of Arrhenius consolidation.

Name

National Academic Infrastructure for Super­computing in Sweden

Legal name

National Academic Infrastructure for Supercomputing in Sweden

Website

https://naiss.se

Company type

Private

Founded year

2023

Operating status

Operating

Headcount range

11–50 employees

Short description

NAISS is Sweden's national academic supercomputing infrastructure, providing free HPC, GPU, storage, and cloud resources to Swedish academic researchers through merit-based allocation, operating multi-petascale systems including TOP500-ranked Arrhenius and dedicated AI clusters.

Ownership category

akta.pro rank

Product category

High-Performance Computing Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518)

SIC

Electronic Computers (3571)

akta.pro primary industry

Network-as-a-Service (NaaS) & On-Demand Connectivity (HDABALAE)

akta.pro secondary industry

Enterprise Storage Arrays (SAN/NAS/Object Storage) (HDACACAB)

Keywords

• High performance computing
• Academic supercomputing infrastructure
• GPU cluster services
• Research data storage
• Cloud computing research

Headquarters

HQ city

Linköping

HQ country

Sweden

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Services

Cost components

Infrastructure, Technology or R&D, Personnel, Operations

Revenue model

1. Government Funding: NAISS is financed by Swedish tax payers through the Swedish Research Council (Vetenskapsrådet) with the objective to strengthen science and research in Sweden. Resources are provided free of charge to researchers at Swedish higher education institutions through open application procedures.
2. EuroHPC Joint Undertaking: NAISS participates in EuroHPC JU initiatives, with Swedish Research Council contributing approximately 3.5% of LUMI funding. Swedish share of LUMI resources (LUMI-C and LUMI-G) is available to Swedish researchers.
3. KAW-Funded Resources: Knut and Alice Wallenberg Foundation (KAW) finances dedicated AI/ML resources including Berzelius supercomputer at NSC, provided to Swedish researchers at no cost for eligible projects.

Pricing tiers

Go-to-market motion3 records

Distribution channels3 records

Marketing channels5 records

Core offering

NAISS provides high-performance computing, GPU-accelerated AI/ML, distributed storage, and OpenStack cloud infrastructure as a national e-infrastructure for Swedish academic researchers. Resources are accessed via the SUPR portal through Small/Medium/Large allocation rounds reviewed by the National Allocations Committee, with systems including Arrhenius, Berzelius, Alvis, Dardel, Bianca, LUMI, Swedish Science Cloud, and Swestore.

Product overview

NAISS (National Academic Infrastructure for Supercomputing in Sweden) is a national e-infrastructure providing HPC resources, storage, and cloud services for Swedish academic research. The portfolio comprises multiple supercomputer systems: Arrhenius (general HPC with CPU/GPU/SENS partitions, plus Disk/Flash storage and PCD cloud service), Dardel (Cray EX system with GraceHopper and GPU partitions), Tetralith, Bianca (sensitive data cluster), Alvis (AI/ML GPU resource), LUMI (EuroHPC share), and Berzelius (KAW-funded AI supercomputer). Storage solutions include Swestore/dCache, Klemming, Berzelius Storage, Mimer (AI storage), and various centre-level storage. The Swedish Science Cloud provides OpenStack IaaS across three regions. Users access resources through SUPR portal and User Area documentation.

Differentiator

Problem solved

Functional benefit

Products and services

• Arrhenius Supercomputer National supercomputer at NAISS ranked #42 on TOP500 (June 2026), featuring CPU partition (192 thin + 20 fat nodes with 256 cores each) and GPU partition (382 GPU nodes with 4x Nvidia GH200 GPUs each). Delivered 66.8 PFLOPS benchmark performance.
• Alvis GPU Cluster National GPU resource dedicated to AI and machine learning research at C3SE/Chalmers, with combined NVIDIA T4, V100, A100, and A40 GPU inventory across Phase 1 and Phase 2 deployments.
• Dardel Cray EX System General purpose Cray EX supercomputer from HPE at PDC (KTH), based on AMD EPYC processors with Lustre storage and Slingshot HPC Ethernet interconnect. Scheduled for retirement January 2027.
• Bianca Sensitive Data Cluster Research system dedicated to analysing sensitive personal data at UPPMAX, providing 4480 cores across dual-CPU nodes with NVIDIA A100 GPU nodes for secure sensitive-data computing.
• Berzelius NVIDIA DGX SuperPOD KAW-funded AI/ML supercomputer at NSC featuring Berzelius Ampere (94 DGX-A100 nodes) and Berzelius Hopper (16 DGX-H200 nodes) connected via NVIDIA Mellanox HDR interconnect. Used for Swedish LLM training with WASP.
• LUMI EuroHPC Consortium Access Swedish share of the EuroHPC LUMI (Large Unified Modern Infrastructure) supercomputer in Finland, including CPU partition (LUMI-C, AMD EPYC) and GPU partition (LUMI-G, AMD Instinct GPUs).
• Swedish Science Cloud (SSC) Geographically distributed OpenStack cloud Infrastructure-as-a-Service across three regions (NORTH-1 Umeå, EAST-1 Uppsala, WEST-1 Chalmers) offering instance types from ssc.tiny to ssc.xlarge with GPU options including A100 variants.
• Swestore/dCache Storage National research data storage infrastructure for active research data, distributed across NAISS centres with dual geographic copies and RAID-6 redundancy, accessible via gridftp, webdav, and NFSv4.1.
• Tetralith General purpose HPC resource at NSC (Linköping University) with 1844 Intel Xeon Gold servers and 170 NVIDIA Tesla T4 GPU nodes, interconnected via 100 Gbit/s Intel Omni-Path. Scheduled for retirement July 2026.
• Kebnekaise HPC Resource Heterogeneous computing resource at HPC2N (Umeå University) with mix of Intel Xeon Gold, AMD EPYC 7763/9754 processors and varied GPU configurations (V100, A100, MI100, A6000, L40s, H100).
• COSMOS Local compute resource at LUNARC for Lund University with 182 AMD 7413 nodes (48 cores, 256 GB RAM), 6 NVIDIA A100 nodes, and 6 NVIDIA A40 GPU nodes.
• Pelle Uppsala University compute cluster installed late 2024 with 48-core standard nodes and GPU configurations including NVIDIA T4, L40s, and H100.
• Mimer AI Storage All-flash storage system at C3SE dedicated for AI/ML workloads attached to Alvis and Vera clusters, providing 0.6 PB all-flash tier plus 7 PB Ceph bulk storage.
• Klemming Project Storage Lustre parallel file system project storage attached to Dardel at PDC, optimized for parallel access with 12 PB total capacity.
• Arrhenius SENS Compute Cluster system within Arrhenius for sensitive data analysis with CPU and GPU hardware, suitable for projects requiring increased security level including sensitive personal data such as whole human genomes.
• SUPR Portal (Swedish User and Project Repository) Web-based portal for managing persons, projects, project proposals, and resource allocations, providing the central access point for all NAISS resources via SWAMID/SAML2 federated authentication.

Quantifiable outcome

• Arrhenius ranked #42 on TOP500 list of world's most powerful supercomputers (June 2026)
• +2 more outcomes

Named customers4 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

AI capability6 records

Feature7 records

Partnerships

Nine partnerships are on record, tiered core, strategic and minor.

• WASP (Wallenberg AI, Autonomous Systems and Software Program)coreStrategic or Co-development Partner · 20 February 2026Joint project with WASP to train high-quality Swedish large language models (LLM) on the Berzelius supercomputer. Researchers from Linköping University, Chalmers, and KTH Royal Institute of Technology participate. WASP provides strategic direction and collaboration framework for AI research in Sweden.
• Mimer AI FactorycoreStrategic or Co-development Partner · 20 February 2026EuroHPC AI Factory initiative in which NAISS and Swedish Research Council invest. Bull (Atos/Eviden) delivering AI-optimized supercomputer for Mimer AI Factory. Combines with Berzelius for comprehensive Swedish AI infrastructure.
• EuroHPC Joint UndertakingcoreStrategic or Co-development PartnerEuropean High Performance Computing Joint Undertaking providing access to LUMI supercomputer and other EuroHPC resources. Swedish Research Council contributes approximately 3.5% of LUMI funding for Swedish share of resources.
• Linköping University (NSC)coreTechnology or IntegrationNSC (National Supercomputer Centre) at Linköping University hosts major NAISS resources including Arrhenius, Tetralith, Sigma, and Berzelius. Provides technical infrastructure, staff, and support for HPC operations.
• KTH Royal Institute of Technology (PDC)coreTechnology or IntegrationPDC Center for High Performance Computing at KTH hosts Dardel supercomputer (Cray EX system), Dardel-GH (NVIDIA GraceHopper), and Klemming storage. Provides technical infrastructure for NAISS resources.
• Uppsala University (UPPMAX)coreTechnology or IntegrationUPPMAX (Uppsala Multidisciplinary Center for Advanced Computational Science) hosts Bianca sensitive data cluster, Rackham, Pelle, and Swedish Science Cloud EAST-1 region. Specialized in life sciences and sensitive data computing.
• Chalmers University (C3SE)coreTechnology or IntegrationC3SE (Chalmers Centre for Computational Science and Engineering) hosts Alvis GPU cluster for AI/ML research, Vera local cluster, and Swedish Science Cloud WEST-1 region.
• NVIDIAstrategicTechnology or IntegrationNVIDIA provides GPU technology across NAISS infrastructure including Arrhenius (GH200), Berzelius (DGX A100/H200), Alvis (T4/V100/A100), and other clusters. Jeff Hammond, Distinguished Engineer at NVIDIA, serves as keynote speaker at NAISS User Forum 2026.
• CodeRefineryminorStrategic or Co-development PartnerNordic training network teaching essential tools for academic computing including git, collaborative development, reproducible research, and software best practices. NAISS actively participates in CodeRefinery workshops.

Scale indicators8 records

Recent moves6 records

Expansion highlights6 records

Broad incumbents

• PRACE (Partnership for Advanced Computing in Europe): European HPC partnership coordinating access to supercomputing resources across multiple European nations. While NAISS is a national operator, PRACE provides the coordinating framework for cross-border allocation, making it a comparable umbrella infrastructure.

Direct peers

• GENCI: France's national HPC agency, operating as a public interest group providing HPC resources to French academic researchers. Like NAISS, GENCI is government-funded, allocation-based, and a key EuroHPC participant, making it a directly comparable national HPC infrastructure operator.
• Pawsey Supercomputing Centre: Australian national HPC facility providing supercomputing resources to Australian researchers. Comparable to NAISS as a government-funded national HPC center with allocation-based access, multiple systems, and a similar mission on a national scale.
• Barcelona Supercomputing Center (BSC): Spanish national supercomputing center operating the MareNostrum series of supercomputers. Comparable to NAISS as a government-funded national HPC provider focused on academic research, with strong involvement in EuroHPC and similar allocation-based access.
• Digital Research Alliance of Canada (formerly Compute Canada): Canadian national HPC infrastructure operator providing advanced computing resources to Canadian academic researchers. Highly comparable to NAISS in its national mandate, multi-site distributed architecture, and government-funded model.
• Jülich Supercomputing Centre (JSC): German national HPC center at Forschungszentrum Jülich, operating among Europe's most powerful supercomputers. Comparable to NAISS in mission, government funding model, and role in EuroHPC, with a similar multi-system portfolio and academic user base.
• CSC – IT Center for Science: Finnish national HPC center that operates LUMI (one of Europe's pre-eminent supercomputers) and provides compute resources to Finnish academic researchers. CSC is the most direct peer because NAISS accesses LUMI through EuroHPC, and the two organizations share Nordic academic HPC planning.
• CINECA: Italian non-profit consortium providing HPC services to Italian universities and research institutions. Operates Leonardo, one of Europe's top supercomputers. Comparable to NAISS as a publicly-funded, allocation-based national HPC provider serving academic research.
• IT4Innovations National Supercomputing Center: Czech national HPC center providing supercomputing resources to Czech academic researchers. Direct peer to NAISS as a EuroHPC participating national infrastructure operator with government funding and allocation-based access.
• Science and Technology Facilities Council (STFC) Scientific Computing: UK research council providing HPC facilities (Hartree Centre, ARCHER2) to UK academic researchers. Comparable to NAISS as a publicly-funded, national-scale HPC infrastructure provider with allocation-based access and ties to national research priorities.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks7 records

Key highlights7 records

Customer concentration

• GitHubgitlab.naiss.se/hpc/user-documentation/support-docs
• LinkedInlinkedin.com/company/naiss-se

Revenue estimate

Valuation estimate

Number of profiles

Profiles5 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about National Academic Infrastructure for Super­computing in Sweden