Eindhoven University of Technology is expected to be among the first European customers to receive NVIDIA DGX B200 systems. The supercomputer - expected to be operational in the first half of 2025 - will boost TU/e's research in AI and other areas.
Frontrunner in AI science
TU/e already uses several supercomputing facilities, but the need for more powerful, high-performance computing is growing exponentially due to the rise of AI in various research areas.
Patrick Groothuis, Vice President of TU/e, states, "We are proud of this new facility, which will greatly enhance the possibilities for our researchers and students to do AI-based research, like cutting-edge cancer research. This also shows our dedication to being a frontrunner in AI science, which will help attract talent and resources and boost cooperation."
The main user groups of the supercomputer will be researchers of the Eindhoven AI Systems Institute (EAISI) of TU/e, and the TU/e Mathematics & Computer Science department. Professor Wim Nuijten, EAISI Scientific Director: "This AI supercomputer stands as a testament to the collaborative efforts between TU/e departments and institutes. Through careful gathering and analysis of requirements from the EAISI research community, we will realize a cutting-edge facility that will remain at the forefront of technology. This allows us to conduct more advanced AI research with increased efficiency and speed."
Sustainable data center
The initial set-up will comprise four NVIDIA DGX B200 systems, each powered by eight NVIDIA Blackwell GPUs, with the option to scale up as needed. NVIDIA DGX B200 systems deliver a unified platform for develop-to-deploy pipelines across diverse AI workloads. The systems will be installed in an environmentally sustainable data center in Finland. DGX B200 systems are industrial-grade, facilitating possible cooperation with industry, which is one of the key strengths of TU/e. The name of the new TU/e AI supercomputer will be 'SPIKE-1'.
Two-year Technical Refresh
The deal concerns an integrated solution (with both hardware and software) whereby the TU/e Supercomputing Center, in cooperation with NVIDIA, will offer researchers help with AI development projects and research issues. It also includes a two-year technical refresh option to renew the infrastructure and software, leveraging the NVIDIA DGX-Ready Lifecycle Management program. This helps ensure that TU/e Supercomputing Center can stay up to date with the latest advancements in computing and maintain uncompromised performance.
