Re:OpenAI is exploring making its own AI chips and evaluates...
by nightski on Friday October 06, 2023 @11:38AM
(#37792224)
Seeing as the GPU is still the primary method used for training ML models, the jury is still out on that. If/when this materializes maybe that will change, but that is a big "if".
If you have an algorithm that works and need scale, you must vertically integrate to maintain an edge over those using more general compute architectures.