Moreh

Stand: 715
Running AI at scale demands solving hard infrastructure problems — parallelization, disaggregation, cluster scheduling, hardware optimization — before a single token is generated. Today, only a handful of companies have the software to do this, and most of it is locked to a single GPU vendor. Moreh builds the software that removes these barriers. We turn heterogeneous accelerators — NVIDIA, AMD, Tenstorrent, and so on — into unified, high-performance AI clusters so that any organization can run frontier models on the hardware they already have.
Visit website (opens in a new tab)
Loading

Contact Exhibitor