PEAK:AIO

Every enterprise AI prompt triggers a hidden 40,000-token workload that creates massive storage demand in the form of key-value (KV) caches, scaling with concurrent users and often overlooked in infra…

PEAK:AIO and Los Alamos National Laboratory have released PNFS-Lattice, an open-source metadata server for parallel NFS that scales to 1,000 servers and accelerates AI data feeds. The software separat…