This confluence created new requirements for HPC storage because the input/output patterns for both workloads were very different. It is happening right now. Intersect360, an independent analyst firm, found that 63% of HPC users are already running machine learning programs. Hyperion Research predicts that the growth in HPC storage spending by public sector organizations and enterprises over the next three-years will be 57% faster than that for HPC compute. Seymour Cray once stated, "Anyone can make a fast CPU." The trick is to create a fast system. Anyone can build fast file storage when it comes to AI and HPC. It is possible to create a cost-effective, scalable and fast file storage system. This is possible by embedding the most popular parallel file systems in parallel storage products from HPE that are cost-effective.