Nvidia Corp. (NASDAQ:NVDA) CEO Jensen Huang says customers keep ordering more AI chips than the company can supply, even as a steep tech selloff raises doubts about how buyers will finance the industry’s enormous expansion.

New orders are arriving “almost weekly and daily,” Huang told the BG2 podcast in an interview released Thursday.

U.S. semiconductor stocks have lost roughly $1.64 trillion in market value over six sessions, Reuters reported, as investors grow concerned about debt-funded AI infrastructure spending.

Nvidia’s Order Book Keeps Growing

“We keep increasing more supply and the order books keeps going up,” Huang said.

Major cloud providers, newer AI cloud operators and equipment manufacturers have already placed substantial orders. Huang said customers forecast demand at least a year ahead, with the largest planning two years out.

Yet Nvidia is allocating those customers considerably less capacity than they want.

Huang said electricity shortages, unfinished data centers and manufacturing components are also limiting how quickly the company can expand supply.

Wall Street Questions Who Can Pay

Investors, meanwhile, are questioning whether AI companies can generate enough revenue to support their enormous borrowing.

CoreWeave Inc. (NASDAQ:CRWV) has roughly $50 billion in debt, while yields on some of its bonds have approached 16%, signaling growing investor concern about its finances.

Yet the Nvidia-backed cloud provider reported a $104 billion revenue backlog at the end of June.

Nvidia owns approximately 10% of CoreWeave and has agreed to purchase otherwise unsold computing capacity through 2032, highlighting Nvidia’s financial exposure to the industry.

“CoreWeave’s strength is improving every day,” Huang separately told Bloomberg Thursday, saying he saw no need for additional financial support.

In the BG2 interview, Huang pushed back on concerns that OpenAI and Anthropic must generate enough revenue to justify trillions in AI infrastructure spending.

“People misunderstand what’s going on,” he said.

Huang estimated frontier AI laboratories account for roughly 25% to 30% of Nvidia’s computing activity. Other demand comes from conventional computing, startups and companies developing specialized applications.

“They don’t have to be anywhere near a trillion dollars in revenues to be successful,” Huang said.

Supporting his demand argument, Taiwan Semiconductor Manufacturing Co. (NYSE:TSM), Nvidia’s primary chipmaker, reported a 68.7% jump in third-quarter profit Thursday.

On Polymarket, traders assigned roughly 6% odds to a qualifying AI-industry downturn by Dec. 31. The contract requires three severe events within 90 days, not merely a stock correction.

Nvidia shares fell 3.2% Wednesday and traded lower again Thursday.

Image: Shutterstock