How HPE and Danfoss collaborated to build Denmark's most advanced energy-efficient AI supercomputer

Wednesday, May 13, 2026

Liquid cooling enables waste heat reuse in the local energy system

The University of Southern Denmark (SDU), Danfoss, and HPE set out to establish a national AI infrastructure to serve researchers across Denmark, delivering high-performance computing through an energy-efficient supercomputer integrated into the local energy system while reusing waste heat.

The challenge

Danish universities lacked a shared infrastructure capable of handling large data volumes and advanced AI models at scale. Researchers and students were constrained in their ability to tackle complex challenges across institutions, and the country's innovation ecosystem — including startups and spin-out companies — needed a platform that could support rapid development, testing, and scaling of new solutions.

At the same time, sustainability was non-negotiable. For Danfoss — a global leader in energy-efficient technologies — the new facility had to demonstrate that advanced AI infrastructure could be built responsibly, contributing directly to Sønderborg Municipality's ambition to build a fully CO₂ neutral energy system.

Sune T. Baastrup (L), Chief Information Officer at Danfoss, shakes hands with SDU rector Jens Ringsmose (R) at the inauguration of Bitten, Denmark’s new energy-efficient supercomputer in Sønderborg, Denmark.

The solution

Danfoss and SDU collaborated with HPE to design and deliver the underlying infrastructure for what has become one of Denmark's most advanced supercomputers for generative AI workloads. Drawing on Danfoss' deep expertise in energy efficiency and industrial decarbonization, the three partners co-developed an approach that went beyond standard data center thinking. HPE's technology was designed to support advanced liquid cooling with full heat recovery — a solution shaped in close collaboration with Danfoss to ensure seamless integration with the local district heating network.

The resulting system — named Bitten, after one of Danfoss' pioneering leaders — delivers high computational power while acting as a productive component of the regional energy system. Waste heat generated by the supercomputer is captured and reused directly in the local district heating network, rather than being released into the atmosphere.

This project demonstrates what becomes possible when research, industry, and technology providers work together with a shared ambition. This new system is among the most advanced in Denmark for generative AI workloads and represents a significant technological upgrade of the national AI infrastructure.

Carsten Nielsen

Vice President and Managing Director for the Nordic Cluster at HPE

The result

For Danfoss and HPE, the project stands as a replicable model for how AI infrastructure can be built more energy efficiently at a global scale. It proves that the right partnerships between industry leaders and technology providers can turn sustainability ambitions into operational reality.

The supercomputer is now part of Denmark's national research infrastructure. It is accessible to researchers and students at universities throughout Denmark via UCloud — a European, sovereign research cloud platform. UCloud has more than 23,000 users, making it one of the largest of its kind globally. Data, software, and computations remain under national and European control, addressing growing concerns around digital sovereignty.

Beyond academia, startups and spin-out companies will gain access to advanced AI and data analytics capabilities, accelerating their ability to develop and scale new products and services.

We have created a unique supercomputing solution that contributes positively across several areas. It delivers both high computational power and acts as an active part of the energy system — demonstrating that sustainability, technological development, and commercially attractive initiatives can go hand in hand.

Sune Tornbo Baastrup

Chief Information Officer at Danfoss

Together, Danfoss and HPE have shown how AI infrastructure can be designed not only for performance, but also for long-term sustainability. The project demonstrates how closer integration between digital infrastructure and energy systems can help support the growing demands of AI while advancing decarbonization goals.