At the current Siggraph event, Jensen Huang of NVIDIA spoke with Wired for an hour about NVIDIA, RTX, and AI. Among other things, he mentioned that in the future, systems powered by artificial intelligence will require three times as many computers.
The NVIDIA CEO pointed out that the world of artificial intelligence is currently transitioning from its initial phase to what he termed the "enterprise wave." This will be followed by a "physical wave." He clarified that three computers will be necessary: one for creating AI, another for modeling and refining AI, and a third for running the AI itself.
“This is a three-computer problem. You know, the three-body problem, and it’s incredibly complex, and we have built three computers for that.”
Huang was referring to a range of NVIDIA hardware and software packages, starting with the DGX H100 servers for AI creation, Jetson embedded computers for AI simulation, as well as workstations and servers using Omniverse and RTX graphics processors to operate AI.
Do we really need three computers for AI? Even companies that genuinely want to integrate artificial intelligence into their core operations cannot overlook the potential cost and complexity of using and paying for three tiers of NVIDIA products. The hardware manufacturer is currently reaping the most benefits from AI technologies and wishes to continue doing so in the future, so such claims should be viewed critically.
Source: PC Gamer
Comments (0)
There are no comments for now