Nvidia CEO: AI’s Next Leap Will Demand 100 Times More Computing Power

*NVIDIA founder and CEO Jensen Huang says said that company’s percentage of revenue in China has fallen by about half due to the export restrictions. Photo/ Nvidia*

Nvidia CEO Jensen Huang says the next wave of artificial intelligence will require 100 times more computing power than previous models, driven by advances in reasoning-based AI that processes information step by step to determine the best possible answer.

“The amount of computation necessary to do that reasoning process is 100 times more than what we used to do,” Huang told CNBC’s Jon Fortt in an interview Wednesday following Nvidia’s fourth-quarter earnings report.

He pointed to models such as DeepSeek’s R1, OpenAI’s GPT-4, and xAI’s Grok 3 as examples of AI systems that employ this new reasoning approach.

Nvidia reported another blockbuster quarter, exceeding analysts’ expectations across the board. Revenue surged 78% year over year to $39.33 billion, while data center sales—dominated by Nvidia’s AI-focused graphics processing units (GPUs)—jumped 93% to $35.6 billion, accounting for more than 90% of total revenue.

Despite the strong results, Nvidia’s stock has yet to recover from its steepest decline since 2020. The company saw a 17% drop on Jan. 27, sparked by concerns that AI firms like China-based DeepSeek could achieve similar AI performance with significantly lower infrastructure costs.

Huang pushed back against those concerns, arguing that DeepSeek’s work only underscores the growing demand for Nvidia’s chips.

“DeepSeek was fantastic,” Huang said. “It was fantastic because it open-sourced a reasoning model that’s absolutely world class.”

Nvidia has faced increasing export restrictions under the Biden administration, limiting its ability to sell its most advanced chips to China. The company relies on billions of dollars in infrastructure spending from the world’s largest tech firms and has been a key beneficiary of the AI boom, with revenue more than doubling for five consecutive quarters through mid-2024 before growth moderated slightly.

Huang acknowledged that export controls have reduced Nvidia’s revenue from China by about half and cited growing competition from Chinese tech giant Huawei as another challenge in the market.

However, he suggested that developers would likely find ways to circumvent these restrictions through software optimization across different devices, whether supercomputers, personal computers, phones, or gaming consoles.

“Ultimately, software finds a way,” he said. “You ultimately make that software work on whatever system that you’re targeting, and you create great software.”

Huang also emphasized the capabilities of Nvidia’s latest AI chip, the GB200, which is available in the U.S. He noted that the chip can generate AI content 60 times faster than the versions Nvidia is permitted to sell in China under current export regulations.