CEO Nvidia says more advanced and models will keep the chip, center data growth

Ai Bubble? Which balloon is? If you ask the entire Nvidia Jensen Huang, we are in the “new industrial revolution”.

Huang’s Company, of course, makes chips and computer hardware, “Tips and shovels” Ai The golden rush and becomes the largest job in the world asking ai’s growth, balloon or not. Speaking on Wednesday during earning a call as his company reported $ 46.7 billion in the past quarter, he cited a sign that incredible growth Generative artificial intelligence The industry will slow down.

“I think next few years, for sure through a decade, we see really significant growth opportunities,” Huang said.

Compare that with Last comments Sam Altman, who said that he currently believes that investors are currently “excessively certain about AI” (Altman also admitted that he still believes that AI is “the most important thing to happen at a very long time.”)

Huang said his company has “very, very significant forecasts” demand for multiple chips and computers that drive ai, which indicates a hurry for more data is not stopping soon. She guessed that AI infrastructural spending could hit three trillion dollars at $ 4 trillion for a decade. (Gross domestic product now is about $ 30 trillion.)

This means a lot of data centers, which occupy a lot of land and Use great water and energy. These AI factories have become larger and larger in recent years, with significant community impacts around them and greater encumber the US electric. And the growth of different generative AI tools that require even greater energy, it could have done that request even greater.

Do not miss any of our impartial technological content and laboratory criticism. Add the CNET As a preferred Google Source on Chrome.

More powerful and more demanding models

One response to Chatbot doesn’t always mean one fast. The source of increased demand for the computer is that newer AI models that use “reasoning” techniques use much more power for one question. “It’s a long thinking, and that longer thoughts, from producing better answers,” Huang said.

This technique allows the AI ​​model to explore on different sites, try the question repeatedly to get better answers and set different information to one answer together.

Some AI companies offer reasoning as a separate model or as a choice marked with something like “deep thinking.” Openay worked rightly In your GPT-5 dismissalWith a guidance program that decides whether it has shown it easier, a simple model or more intense model of reasoning.

But the reasoning model can require 100 times a computer force or more than what a traditional response to a great language, Huang said. These models together with Agent systems It can perform tasks and robotics models that can deal with visualization and work in the physical world, keep demand for chips, energy and transmission centers for on the rise.

“With each generation, demand only grows,” Huang said.