Scientific and high performance computing are active fields of research, but recent or evensemi-recent developments seldomly find their way to industrial applications: - Scientific software is mostly written by experts for experts. - Modelling an industrial process for simulation purposes is tedious and often manual work. - Computational resources are situated in supercomputing centres, not easily acces- sible to the public. - Deriving insights from simulations usually requires both expert tools and expert knowledge. - Manual data processing is not suitable for harnessing the power of machine learning and AI. Ianus Simulation GmbH addresses all these issues with its StrömungsRaum® framework and products for Computational Fluid Dynamics (CFD) applications, including mutable digital twins, automation, predictive surrogate models and automated optimisation [1]. We will demonstrate how robust and automated preprocessing enables non-expert users to use cutting edge scientific software on supercomputers, derive knowledge from automatically generated reports and how AI-based assistance systems will increase the customer’s productivity [2]. A special focus will be bringing exascale computing to the industry by a SCALEXA collaborative research project. REFERENCES [1] M. Geveler and S. Turek. Fundamentals of a numerical cloud computing for applied sciences – preparing cloud computing for “simulation–as–a–service“. Technical report, Fakultät für Mathematik, TU Dortmund, January 2017. Ergebnisberichte des Instituts für Angewandte Mathematik, Nummer 555.2 [2] H. Ruelmann, M. Geveler, D. Ribbrock, P. Zajac, and S. Turek. Basic machine learning approaches for the acceleration of PDE simulations and realization in the feat3 software. In F. Vermolen and C. Vuik, editors, Numerical Mathematics and Advanced Applications Enumath 2019, number 139 in Lecture Notes in Computational Science and Engineering, pages 449–457. Springer, 2020.