Nvidia has been the center of the AI story for the past two years. But are these other AI winners better values today?
Nvidia has been the poster child for the artificial intelligence (AI) revolution thus far, but there will be other AI winners as well.
Now, Nvidia should continue to lead the AI chip market for a while. But the stock has surged 539% over the past three years, and trades at 62 times earnings. Moreover, a slew of competition is heading its way, as the majority of its large customers increasingly look to develop their own in-house accelerators.
In light of Nvidia’s high valuation and incoming wave of competition, the following AI-related stocks may be better values today.
Broadcom
The artificial intelligence buildout is likely to continue strongly for the next few years. And while Nvidia has a near-monopoly on AI chips today, will that continue?
A large majority of AI training happens in the cloud today, and all major cloud providers are now investing heavily in their own custom AI chips. Even large non-cloud tech companies like Meta Platforms are increasingly designing its own custom AI chips in order to lower costs.
While each cloud provider and Meta have solid chip design capabilities, they all need third-party intellectual property to make a significant portion of their custom AI ASICs (application-specific integrated chips). This is where Broadcom (AVGO -2.27%) comes in, providing some key technologies that go into these third-party chips that are produced by Taiwan Semiconductor Manufacturing.
Broadcom already counts Alphabet, Meta Platforms, and TikTok owner ByteDance as customers. Moreover, it’s been rumored that even generative AI model leader OpenAI has recruited Broadcom to design a custom chip for its own models.
Broadcom noted its custom ASIC revenue grew more than 3.5 times, or 350%, year over year last quarter. And CEO Hock Tan said he’d flipped his view about the AI market being dominated by neutral merchant chips like Nvidia’s. Now, he sees custom AI ASICs being taken up at a faster pace, saying, “they [the cloud providers] will all go in that direction totally into ASIC or, as we call it, XPUs, custom silicon.”
Broadcom also has other strong businesses in AI networking chips for Ethernet and VMware software, which the company bought in late 2023. These high-growth businesses are now set to make up more than 50% of Broadcom’s revenue in short order, which could lead to a growth acceleration.
Meanwhile, Broadcom’s diverse business across many types of chips and software give it ample growth opportunities to expand its empire where it sees fit. Since management has a growth-by-acquisition strategy, that means Broadcom has lots of options as to where it can expand its increasingly AI-focused empire. With a growing 1.2% dividend as a kicker, Broadcom looks like a solid value even after its recent run.
Image source: Getty Images.
Lam Research
No matter if Nvidia leads or other competitors emerge in the AI chip space, next-generation AI chips and memory will all be produced by the same handful of semiconductor equipment companies. Of all these competitively advantaged, high-profit equipment stocks, Lam Research (LRCX 1.01%) looks like a solid bet right now after its recent 10-for-1 stock split and its 15% dividend increase.
After surging earlier this year, Lam sold off with other chip names over the summer. Shares currently reside 29% below their highs set back at last July, and at just 23 times earnings estimates for the fiscal year ending in June 2025.
That makes Lam one of the cheaper semicap equipment companies, despite the fact that it’s primed for strong earnings growth. Lam specializes in etch and deposition machines that enable the stacking of chip transistors and other chip elements in a vertical fashion. Until recently, the NAND flash industry has made widespread use of verticalization to stack NAND modules on top of each other; while the NAND business became vastly oversupplied in the last few years, leading-edge logic chips, like AI GPUs and CPUs, as well as DRAM memory, are just beginning to use more vertical structures now.
Lam’s machines help DRAM makers produce the vertical through-silicon vias (TSV) in high bandwidth memory (HBM) for AI applications, and should see increasing capital spend. HBM itself is also a larger die size, meaning there are fewer dies per wafer, which likely means more wafer expansions and more equipment needed for DRAM chipmakers.
Meanwhile, Lam’s atomic-layer deposition tools are required for gate-all-around transistors, the new type of transistor architecture all logic chipmakers are just beginning to use at the upcoming 2nm node next year. Lam expects an extra $1 billion in revenue from gate-all-around nodes this year, and strong growth above that into next year.
Next year, as gate-all-around and HBM kicks into high gear, Lam’s NAND flash segment should also see incremental growth coming off just about the worst downturn in NAND flash history. With NAND equipment investment having nowhere to go but up and its AI-focused segments primed to take off, Lam looks to be a solid growth story for the next couple of years at least.
Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool’s board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s board of directors. Billy Duberstein and/or his clients have positions in Alphabet, Broadcom, Lam Research, Meta Platforms, and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Alphabet, Lam Research, Meta Platforms, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.
