Nvidia Ceo Jensen Huang Says The New Chatgpt Needs '100 Times More' Of His Company's Ai Chips

There's a steep costs to reasoning AI, aliases AI that tin proviso a step-by-step breakdown of its thought process connected pinch a response.

Nvidia CEO Jensen Huang told CNBC connected Wednesday that next-generation AI models tin of reasoning done problems earlier arriving astatine an answer, specified arsenic ChatGPT 4o and Grok 3, require "100 times more" computing power, aliases arsenic overmuch arsenic 100 times nan number of Nvidia's AI chips than older AI models.

This results successful accrued petition and higher income for Nvidia's AI chips, which discarded for up to $70,000 each, to support those reasoning capabilities. Nvidia holds betwixt 70% to 95% of nan marketplace for AI chips.

"We're conscionable astatine nan opening of nan reasoning AI era," Huang told CNBC pursuing nan AI chipmaker's fourth-quarter nett report. "All of these reasoning AI models now petition a batch overmuch compute than what we're utilized to."

Related: DeepSeek AI Cost Less Than $6 Million to Develop. Here's Why Meta and Microsoft Are Justifying Spending Billions.

In nan interview, Huang touted precocious petition for Nvidia's Blackwell AI chip, which had revenues of $11 cardinal successful nan 4th quarter, nan first 4th it was out. Nvidia's accusation halfway revenues, which powerfulness AI systems for companies for illustration Meta, Microsoft, and Amazon, almost doubled from past twelvemonth to scope $35.6 cardinal — and Huang says nan maturation is group to continue.

"This year's superior finance for accusation centers is truthful overmuch greater than past year's," Huang told CNBC.

Meta and Microsoft, which are immoderate of Nvidia's biggest clients, personification already committed billions of dollars this twelvemonth to spending connected AI infrastructure for illustration accusation centers. Meta stated it would locomotion up to $65 cardinal connected AI while Microsoft said it would locomotion $80 billion.

Nvidia CEO Jensen Huang. Photo by Patrick T. Fallon / AFP via Getty Images

Nvidia deed Wall Street estimates connected Wednesday pinch its latest nett report. The AI elephantine reported that quarterly gross accrued 78% to $39.33 cardinal compared to a twelvemonth earlier. Revenue was higher than Wall Street estimates of $38.05 billion.

In nan nett call, Nvidia's main financial serviceman Colette Kress reiterated Huang's relationship astir nan accrued computing power, and corresponding Nvidia AI chips, required to constituent reasoning AI.

"Long-thinking, reasoning AI tin require 100 times overmuch compute per task compared to one-shot inferences," Kress told investors.

Nvidia is nan 2nd biggest institution successful nan world, pinch a marketplace headdress of complete $3.1 trillion astatine nan clip of writing.

Related: 'Everybody Wants to Be First': Nvidia CEO Says Demand for Its Blackwell AI Chip Is 'Insane'