Home Artificial Intelligence Jensen Huang “Even when other firms release AI chips without spending a dime, they won't give you the option to catch NVIDIA.”

Jensen Huang “Even when other firms release AI chips without spending a dime, they won't give you the option to catch NVIDIA.”

0
Jensen Huang “Even when other firms release AI chips without spending a dime, they won't give you the option to catch NVIDIA.”

(Photo = Shutterstock)

Nvidia CEO Jensen Huang asserted that even when competitors release artificial intelligence (AI) chips without spending a dime, they’ll not give you the option to beat GPUs. He cited the overwhelming efficiency of GPUs not only by way of chip price but additionally overall costs, including operating costs.

Tom's Hardware reported on the tenth (local time) that CEO Jensen Huang said that data center operating costs are more vital than GPU prices in a conversation with Stanford University Professor Emeritus John Shoven on the '2024 SIEPR Economic Summit.

In response to this, Professor Shoven asked the query, “Many firms are making inexpensive chips to meet up with Nvidia. Are there any preparations?”

Then, CEO Hwang said, “Even when competitors make free chips, they’ll not give you the option to beat NVIDIA,” and explained that GPU price shouldn’t be very vital when considering the full cost of ownership (TCO) of a knowledge center.

TCO refers to the associated fee of constructing infrastructure and all costs incurred in operating and managing it. What CEO Hwang said signifies that for the reason that efficiency of the GPU is overwhelmingly good, it greater than offsets the high price.

There’s a robust movement by firms reminiscent of AMD and Intel, in addition to cloud firms reminiscent of Google, Microsoft, and Amazon, to exchange GPUs by producing their very own AI chips. Attempts are also being made to exchange 'CUDA', a software system that supports hardware in addition to GPU.

Nevertheless, CEO Hwang announced that he wouldn’t only cooperate with other firms but additionally reveal details in regards to the next chip upfront. That is attributable to confidence that although other firms could make chips optimized for algorithms in specific fields, it should not be easy to make general-purpose and highly efficient chips like GPUs.

He also said, “Individuals who buy and sell chips take into consideration chip prices, and folks who run data centers take into consideration operating costs.” He added, “GPU deployment time, performance, utilization, and adaptability are so excellent that competitors are in a position to purchase the chips.” “Even whether it is released without spending a dime, it should not be competitive,” he said.

In response to this, Tom's Hardware mentioned that competitors would naturally disagree with CEO Hwang's words. Also, many comments say, “It’s annoying.”

Nevertheless, it’s identified that NVIDIA is undoubtedly the world's strongest, and that competitors have to work hard to prove CEO Hwang's words are unsuitable.

Reporter Lim Da-jun ydj@aitimes.com

LEAVE A REPLY

Please enter your comment!
Please enter your name here