Last week, Nvidia (NVDA) stock was the talk of the town, as it vastly beat earnings estimates and guidance, rallying 25% on the next day of trading. The release beat on the reported financials, but what really got people excited was the guidance. Nvidia said on its earnings call that it expected $11 billion in Q2 revenue, $4 billion more than what was expected. Nvidia brought in $6.7 billion in last year’s second quarter, so the projected revenue growth rate is 64%.
It looks like Nvidia is about to regain the kind of growth that it was known for during the tech bull market of 2020 and 2021. Indeed, investors with deep knowledge of both chips and AI may be justified in taking positions in Nvidia now. In a recent article I argued that given Nvidia’s expected growth combined with its extreme valuation, it would perhaps make sense as a hedged long.
For value investors, though, the facts are clear:
Nvidia stock is extremely expensive. Trading at 37.2 times sales, 127 times earnings, and 50 times the best estimate of next year’s earnings, it isn’t the kind of stock that would pass a value screen. If you’re a value investor like me, you’ve probably already said “pass” on NVDA. But that doesn’t mean you have to give up on the AI chip space entirely. There are actually several companies in the space that are rather cheap, including Nvidia’s biggest supplier. In this article I will explore three AI chip stocks that value investors could consider buying in 2023.
Qualcomm
Qualcomm (QCOM) is one of the cheapest chip stocks out there going by multiples. According to Seeking Alpha Quant, it trades at:
-
10.4 times earnings.
-
3 times sales.
-
6.24 times book value.
-
13.82 times operating cash flow.
Compared to the likes of Nvidia, that’s super cheap. Now, of course, this cheapness is for a reason: Qualcomm’s revenue and earnings declined in the most recent quarter. Unlike Nvidia, QCOM did not put out rosy guidance, instead guiding for continued declines in revenue.
Qualcomm will probably not experience the kind of second quarter bonanza that Nvidia anticipates. However, if the company’s AI chip investments pay off, then it could turn its ship around. Qualcomm’s AI investments are a little different than Nvidia’s. It is mainly targeting on-device AI processing rather than data centers. It would make sense that Qualcomm isn’t expecting the kind of growth Nvidia is, because smartphone sales are declining. It pays to consider the long run, though. AI managed to create a lot of buzz around online services like ChatGPT and Bing–perhaps AI-powered smartphones could be the next product category to get a taste of that AI magic.
Taiwan Semiconductor Manufacturing
Taiwan Semiconductor Manufacturing (TSM) is an obvious AI beneficiary that rallied on the same day that Nvidia did. As Nvidia’s manufacturer, it gets an increase in orders whenever Nvidia does.
TSM stock is fairly inexpensive at today’s prices. According to Seeking Alpha Quant, it trades at:
-
15.95 times earnings.
-
6.4 times sales.
-
4.7 times book value.
-
9 times operating cash flow.
For a tech stock in 2022, this is a pretty modest valuation. It’s cheaper than the NASDAQ-100 for example: that index is now at 23.72. So, TSM is cheaper than most big U.S. tech companies.
Alphabet
Alphabet Inc (GOOG) (GOOGL) is a name you might be surprised to see on this list. It is known primarily as a user of chips, rather than a designer. However, if you pay close attention to some of Google’s recent announcements, you will see that it is in fact building AI chips. Specifically, it is building them for in-house use on its own AI projects.
Recently, Google revealed that it had built an “AI supercomputer” that was 1.7 times faster and 1.9 times more power efficient than Nvidia’s A100 systems. It runs Google’s own chipset, dupped TPU (“tensor processing unit”). Google will not be selling these chipsets to other companies, but using them to power its own AI applications. This makes Google a player in the AI chip space–though not a vendor.
At any rate, GOOG stock is very cheap compared to Nvidia. At today’s prices, it trades at 27.9 times earnings, 5.6 times sales, 6 times book value and 17.6 times operating cash flow. Those multiples are from Seeking Alpha Quant, by my own calculations GOOG is at 26 times free cash flow. This certainly is not “rock bottom bargain” territory, but if you compare the company to other AI stocks like Nvidia and Microsoft (NASDAQ:MSFT), which are above 30 times earnings, it looks comparatively cheap.
Honorary mention: ASML Holding
Finally, as an honorary mention, we have ASML Holding (ASML). This stock is too richly valued to be called ‘cheap,’ even in the ‘comparative’ sense of the term I used when defining Google as cheap. However, ASML is worth a mention in any article covering AI, as it is one of the biggest players in chip equipment. As the world’s only supplier of extreme ultraviolet (“EUV”) machines, it is crucial to building advanced AI chips like those being built by Nvidia and Google. At 40 times earnings, it’s no bargain, but it’s closer to the ‘value’ ballpark than the likes of Nvidia. If you’re the kind of value investor who will occasionally buy optically expensive stocks if the competitive position is extremely solid, you may wish to look into ASML. This company is a true monopoly.
Read the full article here