Amazon Web Services took advantage of the Supercomputing ’10 conference to roll out its new GPU Instances. As I write in my post on the new service, it’s a big deal because it further validates AWS’s status as a high-performance computing platform. Why cater to such a relatively small number of users? Because AWS targets a variety of niches, for one, and also because HPC jobs require lots of resources and probably mean a decent profit per user for AWS. In unrelated news, scale-out storage has yet another big day — $25 million in funding for Coraid, and a $2.25 billion price tag for Isilon.