Justin Sullivan
Nvidia Corporation (NASDAQ:NASDAQ:NVDA) just released its fiscal 2024 first quarter earnings. The release beat analyst expectations on revenue as well as on earnings per share (“EPS”). Going into the release, NVIDIA had reached a truly lofty valuation, trading for nearly 29 times sales. Based on post-market trading (NVDA stock is up 20%), it will surpass 30 times sales tomorrow (May 25). Evidently, investors were impressed with Nvidia’s AI chips, and are expecting the company to be the main supplier to a vast chatbot industry.
Indeed, the opportunity that Nvidia is eying may be much bigger than mere chatbots. On March 23, Bill Gates come out saying that the first company to develop an “AI personal assistant” would make search and Amazon.com, Inc. (AMZN) obsolete. The implication here is that this “AI assistant” will be used by as many people as search and Amazon are now.
This could be a huge market. However, as Nvidia’s earnings release showed, it hasn’t been enough to return the company to positive growth. Recently, Meta Platforms, Inc. (META) gave NVIDIA a shot in the arm when it revealed that it was using thousands of Nvidia A100 chips in its AI data center. Things looked bullish for Nvidia at the time. Indeed, today’s earnings release is bullish if you ignore the valuation. It showed a reduction in the magnitude of the negative growth rate, and the potential for future positive growth. Free cash flow nearly doubled!
In this article, I make the case for playing it safe with Nvidia stock, perhaps hedging long bets, and avoiding shorting. Given the stock’s current valuation, it seems like there is a serious risk that longs will experience a prolonged period of negative returns. On the other hand, AI hype is generating enough excitement that shorts could easily get burned. With a stock like NVDA, it’s best to hedge your bets, as I will demonstrate in the ensuing paragraphs.
Earnings Recap
Before getting into my case for avoiding NVDA stock, I need to take a quick look at the earnings just released. The most noteworthy metrics were:
Revenue: $7.19 billion, down 13%, a beat.
Gross margin: 64.6%.
Operating income: $2.1 billion, up 15%.
Net income: $2.04 billion, up 14%, a beat.
Adjusted diluted EPS: $1.09, down 20%.
Free cash flow: $2.64 billion, up 95.5%.
It was a very good quarter. Revenue and earnings growth remained negative, but on the other hand, the buybacks kept coming in strong. I somewhat question whether continued buybacks at today’s valuation are the right thing for NVIDIA’s shareholders, but Wall Street likes to see them, so they should continue being rewarded by the markets.
As for AI specifically, the release showed an 18% increase in data center revenue, thanks to the growing demand for AI chips. In the CFO commentary that accompanied the release, CFO Collette Kress mentioned AI three times, which is surprisingly few for a tech company in 2023. Very frequently these companies mention AI as much as they possibly can in order to drum up interest in their own products. Nevertheless, in this case the actual numbers beat expectations.
Nvidia’s AI Opportunity
There can be no doubt that the AI opportunity in front of Nvidia is massive. Recently Meta Platforms revealed that it was using 16,000 NVIDIA A100 GPUs. At a price of $10,000 a piece, Meta’s A100 order alone cost $160 million! And of course, Microsoft Corporation (MSFT) and Alphabet Inc. (GOOGL) aka Google are well ahead of Meta on AI, using it not only for business processes but also content creation. I was not able to find a precise count for Google’s number of A100 chips, but given that they’re renting out A100 server time to Google Cloud users, the count is likely quite high. Google is a much bigger business than Meta and it does more AI than Meta does, so its number of A100 chips probably exceeds Meta’s.
All of this would tend to argue that Nvidia has a great competitive position in a fast-growing industry (AI, I mean, not semiconductors). However, there’s just one problem.
All of these companies are increasingly looking to build their own GPUs in-house. The story about Meta using 16,000 A100 chips might appear bullish on the surface, but when you read further down the story, you find that Meta actually wants to start designing these chips itself. Apple Inc. (AAPL) is already designing its own chips, they’re built by Taiwan Semiconductor Manufacturing Company Limited (TSM). If more companies go the Apple/Meta route of designing their own AI chips, then the actual AI winner will be TSM, not Nvidia, since it’s the company that all of big tech relies on to build the chips they’ve designed. NVDA’s position here is therefore not unassailable.
Also, there’s the matter of other chip companies competing with Nvidia. Right now, Nvidia is the supplier of choice for big tech companies, but competitors say that they’re closing the gap. For example, Google recently announced that its tensor chips compute faster than the A100. Specifically, the company said that its chips are 1.7 times faster and 1.9 times more power efficient than the A100 is. It’s hard to imagine Google continuing to use A100 systems if these claims are true. That’s a real competitive threat to Nvidia. Not only will Google be able to cease buying from Nvidia, it may even be able to start a new segment selling chips to other companies. That would eat into Nvidia’s margins pretty quickly, though I think given Google’s business model, it would not do that. It would probably prefer to keep the advantage of using tensor chips in house, along with the advantages these chips provide.
Valuation
Now we get to what is by far the worst part of the analysis for Nvidia – far worse than any issues with potential competitors emerging:
Its valuation.
Prior to today’s release coming out, Seeking Alpha Quant had the following multiples on file for NVDA stock:
92 times earnings.
28.3 times sales.
34.24 times book value.
134 times operating cash flow.
All of these multiples are extremely high, higher than what a value investor would pay for a company whose revenue is declining. Also, NVDA’s revenue and earnings went down last quarter, so we know that, as a result of the latest release, the P/E and price/sales ratios are now higher (using today’s closing price as the ‘P’). For example, NVDA had $1.97 in earnings per share over the previous three quarters. That is, $0.88, $0.58 and $0.51 in the quarters prior to the one just announced. The just announced quarter added $1.09 in adjusted EPS. So, we now have $3.06 in TTM adjusted EPS and a 99.8 adjusted P/E ratio. Pretty pricey. And, that ratio will likely go higher tomorrow, as NVDA stock is up 20% after hours.
Risks and Challenges
Nvidia Corporation stock presents a number of risks to both shorts and longs, which is why I’m avoiding it entirely. I think it should decline in price based on fundamentals eventually, but you never know how long AI will keep the gravy train going for the company. So, some risks to investors, short and long, include:
Risk to longs: competition. There has been a trend lately toward tech companies building more and more of their chips in house. After Apple’s big success with its M-series chips, the other big tech companies wanted a piece of the action. This push toward in-house chip design includes GPUs and AI chips. Google, for example, is already saying that its tensor chips beat the A100. Apple has had its own chips for years now. This development is a major threat to Nvidia, particularly if companies like Google decide to start selling their chips to other companies.
Risk to longs: earnings misses. Because Nvidia trades at such high multiples, the potential for extreme reactions to bad earnings releases is very real. Nvidia’s current valuation is due to the fact that people think the company will make a lot of money off of AI in the future. If there is ever any signal that AI hype is dying down, or that Nvidia won’t be the company profiting off it, then its stock may decline precipitously in price. Stocks usually fall in price when they miss on earnings, but the risk is particularly severe when a stock has a high valuation.
Risk to shorts: optimism on the earnings call. Nvidia could easily get its rally going even further by talking about AI a lot on an earnings call, and being very optimistic about its prospects in the space. In the release before this one, Nvidia pulled off a slight beat and only rallied about 4% afterward. It was when the earnings call got underway that the rally pushed well into double digit territory. This time around, the stock is already up 20% after hours, and the earnings call was just getting started when this article was written. I read the CFO’s prepared remarks before the call and they were indeed optimistic. If she impresses Wall Street with her answers to their questions then NVDA stock could rally further still.
As you can see, there are many risks to both short and long Nvidia Corporation traders. Which is why I’m simply avoiding the stock. I just don’t see any good way to play it, except maybe with protective puts. If you do take a position, make sure to hedge it, because Nvidia Corporation trading this year has been very unpredictable.
