Lemurian Labs (www.lemurianlabs.com), based in Toronto and Silicon Valley, is a software company building compilers for GPUs and other processors. They have created an efficient way to make high level AI-related code portable and fast across multiple processors. In addition to AMD and other GPU vendors, every hyperscaler in the AI space is developing custom ASICs for AI workloads, and they will all need an efficient compiler to make their chips fast and easy to use. The Lemurian team has spent decades building compilers at other well known companies.