High Performance Computing & Artificial Intelligence: Practical Use Cases from Industry

The link between high performance computing (HPC) and artificial intelligence (AI) is becoming increasingly important. Training large AI models in particular requires immense computing capacities, which are provided by high-performance supercomputers. At the online seminar of the eco initiative ‘AI in Practice’ on 13 March 2025, experts from business and research highlighted specific use cases that can serve as a cross-industry blueprint for the use of AI in industry.

HPC as an enabler for AI innovations

The High Performance Computing Centre Stuttgart (HLRS) is one of Europe’s leading HPC centres and makes its infrastructure available to both research and industry. GPU-based systems such as ‘Hunter’ enable simulations and AI training to be carried out at previously unattainable speeds. As a member of the Gauss Centre for Supercomputing, HLRS is also one of three federal supercomputing centres and is continuously working to further develop HPC methods in order to open up new fields of application.

In his opening remarks, Hauke Timmermann, Head of Digital Business Models at the eco Association, emphasised the strategic importance of high-performance computing capacities for European companies. Dennis Hoppe (HLRS) presented current developments, including hybrid HPC-quantum computing workflows and the planned ‘AI Factory’, which will provide start-ups and SMEs with dedicated access to AI infrastructure.

Use Cases: From AI-supported automation to real-time analysis

1. Seedbox.ai – AI platform for German industry
The Seedbox.ai platform uses the Hunter supercomputer for training and optimising multilingual AI models. The ‘Kafra’ project focuses on increasing the efficiency of Large Language Models (LLMs) by compressing models and optimising training with large data sets.

2. ARENA2036 – AI and simulation in automotive research
The ARENA2036 innovation network is researching the future of vehicle production. The EU-funded AI-MATTERS project is developing testing and certification processes for the use of AI in manufacturing. Supercomputing enables simulations that digitally optimise real production processes.

3. Festo – automated interaction between robots
Festo combines AI and reinforcement learning to optimise human-robot interaction. An AI-supported gripper technology analyses the shape and texture of objects in real time and adjusts automatically. Supercomputing supports complex training models here.

4. Bosch – AI for optimising manufacturing processes
Bosch uses AI-based analysis methods for predictive maintenance. One example is the optimisation of grinding tools, where machine learning algorithms based on vibration analysis predict the optimal replacement time. This has saved 20% per year in the cost of tools.

Data management and scalability as key challenges

A key topic of the concluding discussion round was the management of large amounts of data. Transferring training data to HPC centres is a challenge for companies. HLRS is therefore planning to set up a dedicated data platform to facilitate access to AI models and open data sets. The possibility of a common industrial data space was also discussed, which could enable the secure and regulatory-compliant exchange of training data.

HPC as a game changer for AI in industry

The online seminar impressively demonstrated how high-performance computing is accelerating the development and implementation of AI in industry. In addition to pure computing power, the linking of AI training infrastructures with industrial processes is the decisive success factor. With projects like the ‘AI Factory’ and European initiatives to promote HPC, high-performance local alternatives to global cloud providers are increasingly emerging.

If you are interested in the further development of AI in industry, consider joining the eco initiative ‘AI in Practice’ and benefit from cooperation with research institutions like HLRS.

Image source: iStock/Fasai Budkaew