This year’s prestigious “Electrifying Ideas Award” from the ZVEI (The German Electrical and Electronic Manufacturers’ Association) goes to Rittal, the GSI Helmholtz Centre for Heavy Ion Research, and to deep-tech company etalytics. These partners have created a German blueprint for the AI centres of the future: Artificial Intelligence (AI) will require more than just algorithms to be effective at scale. It will need data centres that can keep pace with the extreme power density of high-performance chips – ideally, with European sovereignty. The award winners show how this can be achieved with direct liquid cooling and AI-assisted energy optimisation.
As the key voice of the German electrical and digital industry, the ZVEI is presenting its “Electrifying Ideas Award” for the fourth time. A distinguished panel of jury members recognises innovations that set benchmarks, make efficient use of energy, and lead to significant progress for the economy and society. On 20th May, Dorothee Bär, Federal Minister of Research, Technology and Space, presented the award at the ZVEI eSummit in Berlin, the top gathering of the electrical and digital industries.
“Technologies like these are absolutely crucial if we are to build a high-performance AI infrastructure in Germany and Europe. Without innovative cooling concepts, the enormous computing capacities of modern AI systems cannot be operated in an energy-efficient or sustainable manner. I congratulate the team at Rittal, GSI and etalytics on this outstanding achievement, which demonstrates the significant impact that can be generated when industry, research and technological expertise come together,” says Prof. Dr Martin Keller, President of the Helmholtz Association, of which GSI is a member.
“We are very grateful to the jury for awarding us this prize. It draws attention to a pressing issue at just the right time. The expansion of AI-compatible infrastructure in Europe must accelerate considerably,” says Professor Friedhelm Loh, owner and CEO of the Friedhelm Loh Group: “As a world-leading technology company, we have embraced responsibility and have laid the groundwork. Still, there are no surefire recipes for success in implementation. “Our mandate is to work hard to achieve this and set the right priorities in industry, research and politics.”
“This award demonstrates just how important cooperation between research and industry is. Technology transfer paves the way for the findings of cutting-edge research to be put into practical applications,” says Prof. Dr Thomas Nilsson, Scientific Director of GSI and FAIR Dr Katharina Stummeyer, Administrative Director of GSI and FAIR, adds: “Together with Rittal and etalytics, GSI and FAIR have been able to translate innovative and sustainable solutions from basic research into industrial practice – for the benefit of science, industry and society.”
Thomas Weber, Co-Founder & CSO of etalytics, adds: “If we want to remain competitive in Germany in the future, we must make the best possible use of new AI technologies. This is precisely what the collaboration between Rittal, GSI and etalytics has achieved by combining engineering expertise with AI technologies: the context of the cooling system was specifically transferred to the AI application and made usable there.”
Cooling technology paving the way to AI sovereignty
Artificial intelligence holds the promise of rapid progress for research, industry and everyday life – and is of great significance for Germany as a business location. Its growth is also decided in the engine room of digitalisation: In data centres, where new high-performance graphics processing units (GPUs) process enormous amounts of data for AI applications. This is where new cooling and power architectures are needed, as established air-cooling concepts have long since hit their physical limits.
It is precisely here that the nominated project from Germany, with its international expertise, comes into its own: Rittal has devised a direct liquid cooling system using water to directly dissipate heat from the chips. In this way, a computing power of over one million watts in a very small space is made possible – modular, low-maintenance and designed for large-scale 24/7 operation. In collaboration with the GSI, the system was commissioned at the Green IT Cube in Darmstadt and refined to be ready to go into series production. The green GSI data centre offers the ideal environment: state-of-the-art data centre architecture with maximum energy efficiency, high-performance computing for the FAIR particle accelerator, and the Digital Open Lab as a development platform. Furthermore, etalytics complements this concept with AI-supported energy optimisation using digital twins and AI-driven control algorithms, thereby further reducing the carbon footprint.
From bottleneck to key technology with three partners
The result of this collaboration is a niche technology for specialised applications has been transformed into a proven, scalable concept – suitable as a pacesetter for the expansion of sovereign AI infrastructure with expertise and high depth of value creation in Europe.
Rapid, systematic scaling is part of Rittal's international business. The family-owned company from Hesse helps its customers streamline their processes – and not just in the industrial sector – through systems engineering, software and automation. The global player also supplies around 180,000 server racks annually to hyperscalers – the major cloud providers in the USA and Asia. Rittal is also working with them in the Open Compute Project (OCP) on international specifications for the AI factories of the future – for racks and, increasingly, for cooling and power distribution.
Internationally, the GSI is best known for its unique infrastructure for fundamental research. Researchers from all over the world will conduct experiments to gain new insights into the structure of matter and the evolution of the universe at the future FAIR accelerator centre. In addition to research, the GSI uses its “Digital Open Lab” to promote the transfer of technologies and innovations to industry and the digital economy. External partners can leverage the infrastructure and IT expertise for joint development projects and collaborations focused on data centre infrastructure, high-performance computing, and AI.
etalytics ensures transparency regarding key energy consumption figures and optimises them using an AI-powered platform. Having emerged from a research group at TU Darmstadt, the start-up brings DeepTech innovations in data analysis, AI, and energy sciences from the research stage to practical application on an ever-larger scale and for ever-larger companies. By using digital twins of the system, AI improves energy efficiency where AI applications are deployed.