With the Shark A.I. series, Bressner has GPU servers in its portfolio that have been specially developed for artificial intelligence requirements, such as those used in facial recognition, robotics, analytics, healthcare or smart city applications. The GPU server is freely configurable as an edge platform and, according to the provider, can significantly increase flexibility and agility in the cloud. Multiple topologies and bandwidths are supported. Different head nodes can be implemented and the number of GPUs per virtual machine (VM) can be freely selected. The GPU server supports both single- and dual-root complexes for various AI applications. For deep learning applications, a single-root complex can use all GPU clusters to focus on large learning processes and the CPU for small tasks. For machine learning processes, a dual-root complex can assign more tasks to CPUs and arrange less distributed learning processes between GPUs. As the various applications in Edge High Performance Computing require individual platform solutions, the company supports users with customer-specific developments and adapts the CPU, RAM and GPU to the respective requirements. The possible systems support both Nvidia and AMD graphics cards, with up to ten slots possible depending on the model.