Dion from @general_compute showing off his glorious beard standing next to our next rack to take delivery of the @SambaNovaAI SN50. This rack is a beast, it's a turbo charger decode machine that enables much faster inference than GPUs on really big models - but it also has enough prefill compute to be both fast + affordable. I think we were the first inference company to truly see the potential in this rack and bet big on it. Can't wait to plug them in and put into production in early 2027. Congrats @aton2006 @RodrigoLiang on this massive milestone. Keep grinding !