A Nvidia-powered AI supercomputer developed for some of Sweden’s largest companies was officially unveiled in Stockholm on Thursday, 15 October. Named Sferical One, the system is designed to provide shared computing capacity for artificial intelligence development across several strategic industries, including pharmaceuticals, telecommunications and defence.
The project is operated by Sferical AI, a company founded by AstraZeneca, Ericsson, defence manufacturer Saab, banking group SEB and Wallenberg Investments. Industrial engineering group SKF participates as a strategic partner. The initiative brings together businesses associated with Sweden’s influential Wallenberg investment network, as reported by SVT.
Ebba Busch, Sweden’s Minister for Energy, Business and Industry and Deputy Prime Minister, attended the launch alongside Nvidia founder and CEO Jensen Huang and senior executives from the participating companies.
According to Sferical AI’s official announcement, the supercomputer had already been operating since late summer 2026. The October ceremony therefore marked its formal presentation rather than the beginning of operations.
Advanced computing power shared between companies
Sferical One is based on Nvidia’s DGX GB300 technology. Its infrastructure consists of 16 racks containing 1,152 graphics processing units (GPUs), according to Nvidia’s technical overview.
GPUs are specialised processors capable of performing large numbers of calculations simultaneously, making them particularly important for training advanced AI models and running complex simulations.
Rather than developing separate supercomputing facilities, the participating companies share investment, infrastructure and technical expertise. The arrangement is intended to provide greater computing capacity while reducing the need for each business to build an equivalent system independently.
The consortium also places particular emphasis on control over sensitive data and AI development. According to SEB, the infrastructure allows participating businesses to keep their respective computing resources and data separate, even while sharing the overall system.
Sferical AI chief executive Jenny Nordlöw described the platform’s role:
“Sferical One enables the companies to train and run AI models at a larger scale on a sovereign platform.”
The company also plans to develop services using Nvidia’s Nemotron family of open AI models, which can be adapted to individual business requirements and operated in controlled environments.
What the AI supercomputer could be used for
The participating companies are exploring applications across several industries, although the potential benefits remain to be demonstrated on a wider scale.
AstraZeneca is examining applications in pharmaceutical research and drug discovery, including the analysis of complex biological processes. Ericsson is working on AI models for telecommunications networks, while Saab is investigating applications in defence technology and advanced simulations.
The financial sector is another area of interest. SEB has already begun using the infrastructure for selected business applications, with plans to expand access to AI tools for analysing information, developing software and supporting internal operations.
Meanwhile, SKF is exploring industrial uses, including predictive maintenance, quality control and research and development.
These activities represent different stages of exploration and implementation, rather than proven improvements in scientific discoveries or industrial productivity.
A strategic investment in Swedish AI capabilities
Beyond its technical specifications, Sferical One represents an attempt by major Swedish businesses to strengthen their position in the development of artificial intelligence.
The initiative combines access to advanced computing power with greater domestic control over sensitive information, although it continues to rely on technology supplied by the American company Nvidia.
The project also reflects the importance of AI infrastructure for Europe’s industrial competitiveness, particularly in sectors where access to computing capacity and control over proprietary data are increasingly significant.
For Sweden, the consortium establishes a shared technological resource connecting companies in several strategically important industries, including defence. Its longer-term significance will depend on whether the applications now being developed translate into measurable advances in research, services and industrial production.
