Investors want this number to go down to demonstrate that the use of AI is spreading across the economy.

The company’s biggest moneymaker is its accelerator chip, which helps develop AI models by bombarding them with data. Since OpenAI’s ChatGPT chatbot debuted in 2022, a frenzy of artificial intelligence services has created an insatiable demand for the product.

Other recent earnings reports have provided strong signals for AI. Nvidia customers, including Microsoft, Amazon and Meta Platforms Inc., have reaffirmed their commitment to invest heavily in AI infrastructure.

Nvidia appears to be staying ahead of the competition by accelerating the pace of innovation. That includes a commitment to update the lineup annually. With Blackwell, it has a new collection of chips that are faster and have an improved ability to connect to other semiconductors.

But production issues have delayed Blackwell’s rollout. For now, Nvidia cannot fulfill all the orders it receives, the company says. After production improves, supplies will be plentiful, according to Huang.

Nvidia warned again on Wednesday of supply restrictions for Blackwell.

“Critical questions surrounding Blackwell’s production levels and customer concentration remain key concerns,” Emarketer analyst Jacob Bourne said in a note. “There is little room for implementation errors in 2025.”

Nvidia has missed analysts’ quarterly revenue estimates only once in the past five years. And it has exceeded expectations by as much as 20% in recent periods, setting the bar for its performance high.

Its data center division alone now has more revenue than its two closest rivals, Intel Corp. and Advanced Micro Devices Inc., in the aggregate. Net profit this year is on track to surpass the revenue of Intel, a company that for decades was the largest company in the chip industry.

Nvidia made its name selling graphics processors, but discovered that the technology also has applications for AI. The chips help software models during the training process, as they learn to recognize and respond to real-world input. Nvidia’s components are also used in systems that run its software, a stage known as inference, and support services such as ChatGPT.

The Santa Clara, California-based company has quickly expanded its product range to include networking, software and services, as well as fully embedded computing systems. Huang travels the world to lobby for wider adoption of his technology and tries to spread its use by companies and government agencies.

(Updates shares in sixth paragraph.)

Most read from Bloomberg Businessweek

©2024 BloombergLP