NVIDIA recently held an “RTX for Windows AI” event to demonstrate how great its GeForce RTX GPUs are for not only gaming, but also AI purposes, and with it comes a slide that shows how RTX-powered systems are on a completely different, “premium” level, relegating traditional AI PCs (e.g., those powered by AMD, Apple, Intel, and/or Qualcomm chips with NPUs) to what the company can only describe as “basic.” One of the major differentiations lies with their TOPS performance; NVIDIA boasts of potential numbers greater than 1,300 for RTX, while traditional NPUs top out at 45.
Why GeForce RTX is a better choice, per NVIDIA:
More slides that show off RTX’s relevance in the AI space:
Benchlife.info noted in its coverage:
…the necessity of equipped with NPU processors is now widely discussed, and the computing power of known products (whether officially launched or not) is between 10 and 45 TOPS. NVIDIA believes that this can only be used to handle very basic AI operations, such as Windows Studio video special effects processing.