Nvidia Corporation (NASDAQ:NVDA) is arguably the most important tech company on earth. It isn’t the largest – that distinction goes to Apple (AAPL) – but it is the one that other tech companies depend on the most. The company develops the A100 and H100 line of graphics processing units (“GPUs”), which software companies use to power their artificial intelligence (“AI”) applications. AI is extremely resource intensive. In addition to a central processing unit (“CPU”), AI servers also require “accelerator chips” for handling high intensity workloads. Nvidia has the market on such chips cornered, with an 80% market share. As long as this remains the case, Nvidia will enjoy high pricing power, because a lack of competitors means that customers have no choice but to take the price you set.
As long as Nvidia enjoys its dominant market position, it will be able to sell vast quantities of H100 chips for very high prices. An H100 chip can cost up to $30,000, and customers are ordering them by the tens of thousands. As long as this remains the case, Nvidia will enjoy high sales, high margins, and probably high growth as well. The company’s Q3 earnings release (expected post-market November 21st) will most likely show high growth in revenue and even higher growth in earnings. I can’t say whether the growth in these categories will actually be above what analysts expect, but the results should be good enough that people who bought Nvidia stock at the start of the year will remain profitable.
It’s less clear that Nvidia stock will deliver results that will satisfy more recent buyers. As we saw from Meta Platforms (META) and Google (GOOG) last week, even a large beat can trigger a selloff in a stock that trades at high multiples. META and GOOG were both close to the NASDAQ-100’s P/E ratio (about 30) at the time of their earnings releases, and sold off on strong earnings. The reason they sold off was because they had weak results within individual segments: Google’s cloud and Meta’s Reality Labs came in below expectations.
Nvidia stock fell last week as its customers’ shares sold off. For the week, its stock fell 1.44%. The decline from the Tuesday peak was 7%. Clearly, investors think that the segment issues that affected Google and Meta will affect Nvidia. And there may be something to that. Although Google’s selloff on strong earnings seemed irrational, weakness in the cloud segment does point to the possibility that AI hype is dying down. If that’s the case, then Nvidia earnings may come in lower than expected. I am not prepared to declare that that will happen, but as last week showed us, there is a significant possibility that Nvidia will sell off after its earnings come out, whether they beat, miss, or are in line.
Does this mean that it’s time to sell Nvidia stock? Not necessarily. If you bought Nvidia many months ago, before it ever reached the $400 level, you may be justified in continuing to hold. However, the stock probably isn’t the best place to deploy fresh cash into today. Nvidia stock is unbelievably expensive, trading at 77 times adjusted earnings, 97 times GAAP earnings, 30 times sales and 36.6 times book value. If it didn’t grow, then the company would have to pay you 30 years’ worth of revenue in order to pay back your investment! Granted, Nvidia is growing–in the second quarter, it grew quite rapidly. However, it’s not clear that the company’s high growth will persist long term. If competitors enter Nvidia’s market, or if software companies start building their own AI chips in-house, then that could cut into Nvidia’s action. In this article I make the case that Nvidia stock is a “hold,” because it probably does not have much downside here, while also not having a whole lot of upside.
To start things off, we need to take a look at Nvidia’s competitive position. This is by far the biggest thing that the company has going for it right now, so it merits a close look.
Nvidia sells graphics cards to various market segments. Its market shares within these segments are as follows:
As you can see, Nvidia has a majority market share in all of its significant markets. No other company is able to produce GPUs as powerful as Nvidia’s products. Therefore, Nvidia will likely enjoy its near-monopoly in AI chips for the remainder of this year, at least.
However, competitors are beginning to emerge. Among others, there are names like:
Qualcomm (QCOM) – sells on-device AI chips, mainly for use in mobile devices. This doesn’t compete directly with Nvidia’s data center oriented offerings, but if QCOM chips lead to a boom in on-device AI applications, they could indirectly harm Nvidia.
So far, none of these potential competitors are cutting in on Nvidia’s action, but signs are beginning to emerge that some of them may become viable.
Additionally, there’s the matter of companies choosing to sidestep Nvidia and build their AI servers in-house. Almost all of the big tech giants have expressed interest in doing this. Google has even made some progress toward it. That company built an AI supercomputer on its own Tensor Processing Unit chips. It said that the computer was faster than systems running on Nvidia hardware.
Taking all these developments together, we can say this: Nvidia is a near-monopoly in AI chips now, but it may not remain one for the long term. It has 80% of the market, but competitors are seeing the money it’s making, and are eager to get in on the action. So, Nvidia’s fast growth and high margins may not last forever.
AI’s Popularity Waning
Another potential issue for Nvidia’s is a waning in the popularity of AI applications. When ChatGPT launched last November, it took the world by storm, rapidly becoming the world’s fastest-growing application. It was later surpassed in that distinction by META’s Threads, but still, it was successful enough that companies seen making big investments in AI rose rapidly in the markets. For example, Microsoft (MSFT) went on a massive rally after it released its Bing Chatbot.
It was a big moment for AI. Now, however, the hype appears to be dying down. We have seen several signs that AI is not quite the game-changer people thought it was. Some of the big ones include:
ChatGPT’s web traffic declining three months in a row.
Bing failing to take market share from Google Search.
Multiple content creators suing AI applications for using their content without permission.
The AI lawsuits are still ongoing. If generative AI turns out to be a minefield of copyright violations, then its perceived benefits to business will diminish, and investment in AI software may slow down. As a result, we could see an “AI winter,” in which the popularity of AI applications declines. Such events have occurred in the past; there’s no reason they couldn’t occur in the future.
The Bottom Line
The bottom line on Nvidia is that it’s a great business, but not one worth an infinite price. To be sure, the company’s growth last quarter was off the charts, and next quarter will probably be strong too. However, the company’s competitors are ramping up their investments, and AI is showing early signs of waning popularity.
There are issues for those who take either side of the Nvidia trade. For those going long, increased competition, a possible AI winter, or even just an unexpectedly sour reaction to a good fiscal Q3 earnings release could spoil the party. For those choosing to short Nvidia, the company’s persistently high growth is a challenge. On the whole, Nvidia stock is a classic “hold”–worth sticking with if you got in substantially below today’s price, but not worth buying or shorting today.