Computational fluid dynamics (CFD) traditionally has been one of the most compute-intensive disciplines in computational sciences. Typically more than half of compute resources of the Tier 0 computing centers around world is devoted to fundamental-research problems using CFD. CFD has become a standard tool in daily industrial work, disaster prediction and prevention, entertainment industries. In order to enable more powerful predictive tools, better efficiency, and higher accuracy, CFD always has inspired research in applied mathematics and in computer science. CFD has faced and mastered disruptive changes in computer hardware architecture and incorporated dramatically improved algorithms. While doing so important new concepts for algorithm development such as co-design have been developed, which make CFD more resilient to future disruptive changes. The mini symposium aims at gathering emergent concepts for next-generation CFD from a wide range of directions, and providing a forum for mutual assessment and discussion. While the inclusion of data-driven surrogates and of full substitutes for classical predictive simulators is already a fast progressing subject, systematic replacement of numerical approximation of ground truth mathematical models by algorithmic embodiments is still at its infancy. The same applies to hardware hybridization, where acceleration by GPU and TPU are well within the current scope, with quantum-acceleration still being a speculative entity. Specifically we are interested in contributions on: • Performance-portability and exascale computing in CFD • Data-driven surrogates for generative or predictive tasks in CFD • Quantum-inspired algorithms for accelerating CFD • Automatic differentiation and inverse problems in CFD • Advanced hybrid numerical methods and turbulence models for CFD