Last week, Nvidia (NVDA -1.85%) reported better-than-expected results for the third quarter of fiscal 2024 (ended Oct. 29). The graphics processing unit (GPU) specialist's growth was driven by voracious demand for its products that enable companies and other entities to develop and deploy artificial intelligence (AI) technology.

In fiscal Q3, Nvidia's revenue surged 206% and its adjusted earnings per share (EPS) rocketed 593% year over year. Both results raced by Wall Street's estimates, as did the company's top- and bottom-line guidance for the fiscal fourth quarter. For Q4, management expects year-over-year revenue and adjusted EPS growth of 231% and 407%, respectively.

Earnings releases tell only part of the story. Management often shares information about a company's prospects on the quarterly earnings calls. Indeed, on Nvidia's Q3 earnings call, CEO Jensen Huang and CFO Colette Kress discussed two new markets they view as multibillion-dollar opportunities for the company.

1. Sovereign AI infrastructure

From CFO Kress' remarks:

Many countries are awakening to the need to invest in sovereign AI infrastructure to support economic growth and industrial innovation. With investments in domestic compute capacity, nations can use their own data to train LLMs [large language models] and support their local generative AI ecosystem. ... National investment in compute capacity is a new economic imperative and serving the sovereign AI infrastructure market represents a multibillion-dollar opportunity over the next few years. [Emphasis mine.]

First, a definition: "Sovereign entities" are those that are independent and have total or at least considerable control within their borders. This includes many nations, U.S. states, and the European Union. In the U.S., Native American tribes are another example.

Nvidia has made some good inroads with the sovereign AI infrastructure market. Notably, in September, it announced that it was working with the government of India and the country's largest tech companies, including Infosys, Reliance Industries, and Tata Group, to improve India's sovereign AI infrastructure.

And earlier this month, French cloud service provider Scaleway, a subsidiary of telecom company Iliad Group, announced that it's building a European regional AI cloud service based on Nvidia's powerful H100 GPUs, Nvidia's InfiniBand networking product, and Nvidia AI Enterprise software. Scaleway provides sovereign infrastructure that "ensures access and compliance with EU data protection laws," Nvidia said in November on its blog.

CEO Huang added that the "number of sovereign AI clouds that are being built is really quite significant." He said the key reason is that "people realize that they can't afford to export their country's knowledge [data used for AI training] ... for somebody else to then resell AI back to them."

2. Grace Hopper Superchip

From Huang's remarks:

Grace Hopper is in high-volume production now. ... We are on a very, very fast ramp with our first data center CPU [central processing unit] to a multibillion-dollar product line. ... The capability of Grace Hopper is really quite spectacular. It has the ability to create compute nodes that simultaneously have very fast memory as well as very large memory. [Emphasis mine.]

The just-reported fiscal third quarter was the first quarter that Nvidia generated revenue from sales of its new GH 200 Grace Hopper Superchip. Huang said that the company is expecting a big year next year for this product given the many design wins the company has had.

Nvidia describes the GH200 Grace Hopper Superchip as a "breakthrough accelerated CPU designed from the ground up for giant-scale artificial intelligence and high-performance computing." This superchip combines Nvidia's ARM-based Grace CPU with a Hopper GPU. Basically, it brings together the strengths of the GPU and the CPU.

Kress said that Grace Hopper instances are now available at GPU-specialized cloud service providers and are coming soon to Oracle Cloud. Moreover, in fiscal Q3, several supercomputing entities received shipments of the superchips, including the Los Alamos National Lab in the U.S. and the Swiss National Supercomputing Center.