Nvidia and HPE Team Up for Blue Lion: New Supercomputer in Germany by 2027 | List of hardware components and functions | Gpu vs cpu | Gpu example | Turtles AI

Nvidia and HPE collaborate with Leibniz Supercomputing Centre to build BlueLion, a European supercomputer equipped with VeraRubin processors, accessible to researchers in 2027. In parallel, Jupiter is officially the most powerful machine in Europe and new AI models for climate debut.

Key points:

• Partnership between Nvidia, HewlettPackardEnterprise and LRZ for BlueLion.
• VeraRubin architecture (Rubin GPU + Vera CPU) will increase HPC capabilities up to 30× compared to SuperMUC-NG.
• AI “Climate in a Bottle” for climate forecasts up to 30 years on a kilometer scale.
• Jupiter, already operational at Forschungszentrum Jülich, is now the fastest supercomputer in Europe.

Nvidia and Hewlett Packard Enterprise announced during a conference in Hamburg an agreement with the Leibniz Supercomputing Centre to build the BlueLion supercomputer in Germany, scheduled for early 2027 and equipped with the new VeraRubin architecture: a combination of Rubin GPUs and Vera CPUs, capable of aggregating simulation, data and AI in a single infrastructure with very high bandwidth and low latency, offering up to 30 times the computing power of the current SuperMUC-NG. The system will exploit advanced HPE Cray technologies, direct liquid cooling without fans and heat recovery for local heating, and will be available to European researchers working in fields ranging from climate to physics, from turbulence to AI. In parallel, at the Forschungszentrum Jülich, the Jupiter supercomputer, based on Nvidia chips, has been recognized as the fastest system in Europe, signaling the desire of the community institutions to maintain a competitive profile with respect to the United States in high-performance computing in the scientific and environmental fields. At the same time, Nvidia launched “Climate in a Bottle,” an AI model that can receive initial conditions such as sea surface temperature and generate detailed climate scenarios at kilometer level for the next 10-30 years, thanks to a synergy between classical physics methods and artificial intelligence. These developments demonstrate the company’s commitment to promoting an accelerated scientific approach, in which precise calculations are complemented by predictive AI models to rapidly probe thousands of atmospheric scenarios in high detail.

This project represents a significant milestone in the European strategy to strengthen supercomputing and AI capabilities applied to scientific research, maintaining a technical and collaborative approach that does not produce anticipations or judgments, but outlines a significant technological progression.