But the current AI boom “has ushered in the largest infrastructure expansion in human history,” Huang said.”We’re a few hundred billion dollars into it now… there are trillions of dollars of infrastructure that needs to be built out” in areas such as energy, cloud computing and electronics, he added.
Nvidia dominates spending on the processors needed to train and operate the large language models (LLMs) behind chatbots like OpenAI’s ChatGPT or Google Gemini.
Sales of its graphics processing units (GPUs) – originally developed for 3D gaming – took the company’s market capitalization to over $5 trillion in October, although this figure has since fallen by more than $600 billion.
LLM developers like OpenAI are directing much of the massive investment they’ve received into Nvidia’s products and rushing to build GPU-filled data centers to meet the expected demand for AI services. Huang on Wednesday dismissed warnings from some observers that the spending spree is increasingly looking like a bubble that could soon burst.
“The AI bubble is happening because the investments are big. And the investments are big because we have to build the infrastructure necessary for all the layers of AI above,” he said.
“I think the opportunity is really quite extraordinary.”
Huang also sought to allay fears that AI would destroy large numbers of jobs as its use trickles down to workplaces, saying instead that it would create new forms of work.
Microsoft CEO Satya Nadella struck a more cautious tone on Tuesday, saying widespread adoption of AI was needed to avoid an industry crash.
“To avoid this being a bubble by definition, the benefits of this need to be much more evenly distributed,” he said.
Nadella was nevertheless convinced that this is a technology that will in fact build on the rails of cloud and mobile, and spread more quickly, driving “economic growth around the world”, he added.
#Nvidia #boss #sees #trillions #spending #ahead