“Everything sold out”: Nvidia’s Huang rejects AI fatigue concerns
The CEO says demand for Blackwell and Hopper chips proves boom is far from over, even as forecasts fall short of sky-high expectations.
Nvidia CEO Jensen Huang on Wednesday downplayed concerns about an end to the spending boom on artificial intelligence chips, projecting opportunities will expand into a multi-trillion-dollar market over the next five years.
Huang sought to reassure investors unsettled by signs of slowing growth at the chipmaker that has been at the center of the AI investment frenzy. Earlier in the day, Nvidia forecast third-quarter revenue in line with analyst expectations, but short of the lofty projections that have driven its share price up roughly one-third this year.
The founder and CEO’s bullish outlook contrasted with recent signals of fatigue in AI-focused stocks and warnings from industry leaders about overheated investor enthusiasm.
“A new industrial revolution has started. The AI race is on,” Huang said. “We see $3 trillion to $4 trillion in AI infrastructure spend by the end of the decade.”
Expectations of sustained demand from Big Tech, data center operators known as hyperscalers, and buyers in China have helped push up Nvidia’s shares.
“The mega caps are the ones propelling a lot of the capex that Nvidia is benefiting from. But obviously Nvidia is still growing, still able to sell,” said Matt Orton, head of advisory solutions at Raymond James Investment Management. “If anything, this just highlights that there’s a lot of durability to this (AI) trade. The businesses of these hyperscalers can continue to accelerate, and you’re not seeing any sort of sign of a slowdown reflected in Nvidia’s results.”
While Nvidia shares have outpaced a roughly 10% gain in the broader market, AI stocks more broadly have shown signs of fatigue. OpenAI CEO Sam Altman warned this month that investors may be “overexcited” about AI.
On Wednesday, Huang appeared unperturbed. “The more you buy, the more you grow,” he said, arguing that Nvidia’s technological advances allow customers to process greater amounts of data while using less energy. “The buzz is: everything sold out.”
As an example, Nvidia said a customer outside China bought $650 million worth of its H20 reduced-capability chip aimed at the Chinese market in the latest quarter.
Huang based his forecast partly on an expected $600 billion in data center capital spending this year from major customers such as Microsoft and Amazon. For a data center costing as much as $60 billion, he said, Nvidia could capture about $35 billion.
His remarks contrasted with Nvidia’s muted third-quarter sales forecast of about $54 billion, just ahead of the $53.14 billion average analyst estimate compiled by LSEG.
Related articles:
Still, Nvidia and Huang see little reason for AI chip profit growth to slow, as the company’s second-quarter net income already surpassed the fiscal third-quarter profit of Big Tech peer Apple.
Demand remains strong for its high-end Blackwell chips, largely sold out through 2026 based on customer forecasts. Its earlier-generation Hopper processors are also being rapidly snapped up.
“When you have something that is new, and it’s growing as fast as it is, and with all of the huge capex announcements from the hyperscalers, it’s evidence that we’re in the early stages of the AI boom,” said Thomas Martin, a portfolio manager at Globalt Investments.
