This suggests
"an opportunity to translate success in HPC to boosting the performance of applications based on Microsoft platforms,"
No it doesn't.
It suggests that where performance and indeed price/performance is more critical than the variety of desktop-style application availability, Linux remains the obvious choice, and Windows remains largely irrelevant.
For years, comparing the performance of similar hardware running different OSes has been made as difficult as possible by the big box builders who don't want to upset their sweetheart deals with MS. But whenever I've found a comparable benchmark running on comparable hardware with the two different OSes, the winner has been entirely predictable, and generally hasn't been Microsoft.