Stuttgart, December 10, 2024 — The EuroHPC Joint Undertaking (EuroHPC JU) today announced the creation of a new consortium called HammerHAI, which will establish a new AI Factory in Germany. Coordinated by the High-Performance Computing Center of the University of Stuttgart (HLRS), HammerHAI will address the urgent demand for increased AI capabilities in academic research, start-ups, SME’s, and European industry, as well as in the public sector. The AI Factory will install a secure, AI-optimized supercomputing infrastructure, provide expert support services, and develop solutions to make it easier for researchers and companies of all sizes to access and use powerful AI technologies.
Dec 10, 2024
See all press releases
Bringing together prominent organizations within the German high-performance computing (HPC) and artificial intelligence communities, HammerHAI will quickly build an AI Factory on a solid foundation of existing infrastructure and expertise. Joining HLRS in the project is the Leibniz Supercomputing Center, which like HLRS is also a member of Germany’s national HPC infrastructure in the Gauss Centre for Supercomputing. In addition, the AI Factory will benefit from important contributions by the Gesellschaft für wissenschaftliche Datenverarbeitung Göttingen, the Karlsruhe Institute of Technology, and HPC/AI consultancy SICOS BW.
The project budget of approximately 85 million Euros is co-funded by the EuroHPC JU, the German Federal Ministry of Education and Research, the Baden-Württemberg Ministry of Science, Research, and Art, the Bavarian Ministry of Science and Art, and the Lower Saxony Ministry for Science and Culture. HammerHAI is one of seven European AI Factories announced today by the EuroHPC JU.
“The European and German governments have identified the creation of a robust, native European infrastructure for artificial intelligence as a high priority. HammerHAI will quickly help to address this need,” said Prof. Dr. Michael Resch, Director of HLRS. “At the same time, the project will play an important role in implementing the development strategy at HLRS and the Gauss Centre for Supercomputing as we evolve to support the changing landscape of high-performance computing.”
Currently, scientists and companies across Europe face a number of hurdles that have limited the adoption of artificial intelligence technologies. These include the prohibitive cost of installing AI computing hardware, the shortage of AI expertise, and data security concerns arising from Europe’s current reliance on offshore cloud AI service providers.
By establishing a one-stop shop that addresses the needs of European AI users, HammerHAI aims to overcome these challenges. “Working closely with the European artificial intelligence community, HammerHAI will both deploy new AI-optimized computing infrastructure and develop a comprehensive portfolio of services and technologies,” said HLRS Managing Director Dr. Bastian Koller, who will manage the project. “In this way, the AI factory will make it easier for scientists, startups, and SME’s to seize the opportunities that AI offers for discovery, innovation, automation, and decision making.”
HammerHAI will build on the existing capabilities of Germany’s national high-performance computing infrastructure and procure a new, large-scale AI-optimized supercomputer, which HLRS will operate on behalf of the EuroHPC Joint Undertaking and the project funders. The system will be tailored to the computational requirements of typical AI workloads in industry, including the training and usage of customized large language models, deep learning, and complex data analytics. The new system will also accommodate hybrid workflows that integrate HPC and AI, such as using HPC to generate synthetic datasets for the training of neural networks, or using AI to accelerate computationally demanding simulations through surrogate models.
The goal of HammerHAI is to provide a secure, accessible, and scalable resource that will become a cornerstone for Europe’s AI ecosystem. It will support the development and testing of new applications as well as the migration and evolution of existing solutions that would benefit from the increased performance of an AI-optimized supercomputer. Migration will be facilitated through a cloud-native approach to HPC that offers web-based interfaces, an easy mechanism for data sharing, and the support of virtualisation and containerisation technologies. The system will also integrate pre-packaged AI frameworks like TensorFlow and PyTorch, and offer access to pre-trained AI models and large datasets.
Additionally, AI specialists at HammerHAI will promote the adoption of AI methods by assisting users as they migrate their applications from the cloud to HPC and offering comprehensive end-to-end support for the entire AI life cycle, from model development to the use of AI models in a production environment. Continuing professional education courses within the HammerHAI consortium will also empower application programmers to acquire the skills they need to fully exploit the potential of AI technologies.
Importantly, HammerHAI will provide a secure, Germany-based platform and data repository. It will place strict controls on access to the resulting AI models, the data on which they are based, and the query interfaces through which they are utilized. This will make it easier for European researchers, startups, and SME’s to adhere to compliance requirements concerning data security and the ethical use of AI. A publicly funded national high-performance computing center, HLRS is certified for information security under the ISO 27001 standard and has completed a Level 3 audit under the Trusted Information Security Assessment Exchange (TISAX), a framework used for evaluating data centers with “very high protection needs.”
HammerHAI will be embedded within the German and European AI ecosystems. The project partners are closely integrated into numerous regional and national AI initiatives across Germany, and as coordinator of European HPC networking projects such as EuroCC 2, CASTIEL 2, and FFplus, HLRS will promote the availability of and access to the AI resources that HammerHAI will offer. In this way, the German AI Factory will become an important component of a Europe-wide effort to transform industry and support the public and academic sectors in the AI era.
The High-Performance Computing Center Stuttgart (HLRS) is a research institution affiliated with the University of Stuttgart and a founding member of the Gauss Centre for Supercomputing, the alliance of Germany's three national supercomputing centers. HLRS will lead and manage HammerHAI, and provide computational and storage infrastructure and resources, including the necessary software, access models, and information security measures.
The Leibniz Supercomputing Center (LRZ) of the Bavarian Academy of Sciences and Humanities is also a national supercomputing center within the Gauss Centre for Supercomputing and is the IT service provider for universities in Munich and research organizations throughout Bavaria. As a partner in HammerHAI, LRZ will focus on issues related to user training, provisioning and maintenance of the AI Factory’s core AI software environment, data management and access, and AI trustworthiness.
The Gesellschaft für wissenschaftliche Datenverarbeitung Göttingen (GWDG) is a joint institution of the Georg-August-Universität Göttingen - Stiftung Öffentlichen Rechts and the Max Planck Society, and the central computing facility of the University of Göttingen. Building on expertise gathered through hosting the AI service center KISSKI, the GWDG will interface with other national AI service centers and provide support to clients of the AI Factory, especially with end user analysis and the development of solution architectures.
The Karlsruhe Institute of Technology (KIT) is a founder of the Smart Data Innovation Lab, bringing more than 10 years of knowledge concerning data driven innovation, trustworthy AI and embedded AI, and the use of data-based methods in SMEs. KIT and HLRS currently collaborate in the KI Allianz BW by combining KIT's expertise in AI research and innovation with HLRS's advanced infrastructure to drive the data ecosystem for academia and industry in Baden-Württemberg.
SICOS BW was cofounded by HLRS and KIT to support small and medium-sized enterprises in overcoming the hurdles that stand in the way of using high-performance computing and artificial intelligence. As a partner in HammerHAI it will focus on public relations, communication with users, and community building.
Press release by the EuroHPC Joint Undertaking Selection of the First Seven AI Factories to Drive Europe's Leadership in AI
Press release by the German Federal Ministry for Education and Research Özdemir: AI Factory kommt nach Deutschland – Großer Fortschritt für KI-Infrastruktur in Deutschland und Europa
Press release by the Baden-Württemberg Ministry of Science, Research and Art Stuttgarter Konsortium baut europäische KI-Fabrik
Dr. Bastian Koller Managing Director High-Performance Computing Center Stuttgart Tel.: +49 (0) 711 / 685- 65891 koller(at)hlrs.de
Sophia Honisch Head, Public Relations High-Performance Computing Center Stuttgart Tel.: +49 (0) 711 / 685-68038 honisch(at)hlrs.de
