Artificial intelligence (AI) absolutely dominated the stock market for the past few years, and no company dominated AI quite like Nvidia (NVDA -0.40%). Its advanced chips are the bedrock that enable AI, lining the huge data centers that power ChatGPT and other AI models.
Given its place at the center of the AI boom, the company is seen as a bellwether for the entire sector. That's why all eyes will be on Nvidia's upcoming earnings report on Aug. 27.
Is demand still white-hot? Or are there signs of slowdown? With trillions of dollars of market value at stake, a lot is riding on the results. With this in mind, is now the time to buy this ultimate AI stock?
AI demand is driving Nvidia's GPU sales higher
Nvidia's premier GPUs are the beating heart of AI data centers. Its current Blackwell chips are -- as its earlier Hopper chips were -- far ahead of the competition, driving explosive growth in both revenue and profit. AI GPUs account for about 80% of Nvidia's business, and demand still looks white-hot, but investors will be eager to see if this is indeed true. Are sales still growing rapidly? Is it maintaining its incredible margins?
While we won't know for sure until Aug. 27, all the signs point to yes. Many of its downstream customers have reported strong results and committed to record levels of capital expenditures, much of that earmarked for its GPUs.
Networking could be Nvidia's next growth engine
Most investor attention has been on Nvidia's GPUs, and with good reason: They are the moneymaker for the company. However, a promising source of growth and one that some investors have overlooked is its networking hardware.
Its GPUs process mind-boggling amounts of data, but that data has to travel among the GPUs and servers that make up a data center. Traditional Ethernet-based networking often doesn't cut it.
That's where the company's hardware -- NVLink, InfiniBand, and Spectrum X -- comes in. It will be a major source of revenue, and networking hardware is already its fastest-growing segment.
Last quarter, Nvidia reported a 64% jump quarter over quarter for its networking hardware, reaching $5 billion in sales. If that sort of growth continues, its networking segment could become a much more substantial part of its overall revenue mix and drive its top line even higher.
Image source: Getty Images.
CUDA is key to Nvidia's moat
While its technical advantage on the hardware side can't be overstated, Nvidia's most durable advantage is actually in software. Its CUDA platform is a sort of programming framework that lets developers use chips in parallel for tasks beyond their original intention (computer graphics), like training and running AI models.
Over the years, it has become the foundation upon which the vast majority of AI programs are built. CUDA is proprietary; it's also sticky. This gives Nvidia a wide moat. If a company wants to move to a rival's chips, it would have to rework much of its entire software stack and retrain its engineers -- or even be forced to hire new engineers.
That's an enormously costly and time-consuming process and a large part of what keeps customers locked in its ecosystem, willing to pay a premium. This moat helps protect Nvidia's business even if its current hardware leadership evaporates.
Generative AI still needs to prove it's profitable
While the outlook for Nvidia looks strong, it's worth considering the risks. In the near to medium term, record capital spending from big tech should keep demand high. But the long-term viability of today's AI model is less certain, and there are reasons to be skeptical.
Take OpenAI. It isn't public, but reported financials show rapid revenue growth paired with huge losses. Generative AI is expensive to run and very expensive to train. At present, the business model is just not profitable.
Optimists argue that costs will fall and capabilities will improve dramatically, but if that doesn't happen -- either to the degree it needs to or in a reasonable time frame -- companies like OpenAI and Anthropic could find themselves in hot water. If one or more of these central figures in AI folds, demand for Nvidia's chips would take a huge hit.
I'm not saying by any means that this is going to happen, but it's important for investors to keep a clear view and not get caught up in hype.
Nvidia is still a buy
Nvidia's stock already carries a hefty premium, with much future growth already priced in. Even so, near-term catalysts and strong demand for its market-leading chips make the stock a good investment ahead of Aug. 27, in my view. I expect management has a good chance of beating the high bar it has set.
And over the long run, emerging markets like robotics -- just as dependent on Nvidia's GPUs as generative AI is -- could drive another wave of growth or offset any trouble in generative AI.
