The artificial intelligence (AI) revolution appears to be the real deal rather than hype. After all, ChatGPT was by far the quickest application to reach 100 million users -- in just two months! In a recent interview, Cathy Gao, partner at venture capital firm Sapphire Ventures, anticipated "a future where AI becomes so ubiquitous that companies no longer market themselves as 'AI companies' because they've all become AI companies."
Along with inflation coming down, enthusiasm for AI has driven the Magnificent Seven stocks significantly higher this year. But while the "Mag Seven" have been propelled by leadership in cloud and semiconductors, there are plenty of other crucial companies within the AI ecosystem that will benefit as well.
The following three stocks look like excellent additions to any AI portfolio beyond those popular seven, making for quite a nice top 10.
These AI stocks are cheaper than the Magnificent Seven. Image source: Getty Images.
Taiwan Semiconductor Manufacturing
Whether or not Nvidia or another of its up-and-coming competitors leads the artificial intelligence chip race, Taiwan Semiconductor Manufacturing (TSM 1.26%), known as TSMC, will manufacture all or most of those chips.
TSMC pulled ahead of other foundries around five years ago in terms of chipmaking process technology and has compounded that lead ever since. So all of the world's leading AI companies, from Nvidia to Advanced Micro Devices to the cloud computing giants that design their AI chips in-house all depend on TSMC to manufacture them. In fact, despite having a massive market share already, TSMC increased its foundry market share in the third quarter from 56.4% to 57.9%, according to TrendForce. The next-largest competitor, Samsung, had just 12.4% share.
That is why TSMC was able to easily raise prices and pass along costs during the high-inflation period following the height of the pandemic, which attracted the attention of Warren Buffett -- at least temporarily.
What's also highly appealing is that TSMC trades at a below-market multiple, at just 17 times its forward earnings estimates, and it pays a not insignificant 1.84% dividend, which should grow into the future.
For sure, there are some risks, as there are for any stock in the rapidly changing tech industry. For one, both Samsung and Intel have stated they intend to catch up to TSMC in 2025, when the industry progresses to 2nm chips. In addition, Buffett stated he sold TSMC shortly after buying it due to its location in Taiwan, which China claims as its own territory.
But both of these concerns may be overblown. While some customers may outsource some of their production to other foundries, TSMC's momentum, scale, and manufacturing know-how are likely hard to replicate. So TSMC is likely to be the industry leader years from now as well. Furthermore, TSMC is opening up new fabs in the U.S., Japan, and Europe, somewhat mitigating the geopolitical risk. And TSMC's low multiple also means investors today are already paying a discounted price.
At a high level, TSMC is one of the most important companies in the AI chipmaking ecosystem, making it a value investor's choice in the AI age.
Super Micro Computer
Sort of like TSMC, no matter which chipmaker wins the AI wars, a lot of those chips will be run on Super Micro Computer (SMCI 8.89%) servers.
Super Micro stock has tripled this year, so some may think they've missed the big opportunity. But also like TSMC, the stock only trades around 17 times its fiscal 2024 earnings estimates, with SMCI's fiscal year ending in June.
While server makers are often thought of as commoditized companies that assemble proprietary chips and networking gear, the demands of artificial intelligence are putting a greater emphasis on server design.
That's where Super Micro shines. Its innovative "building block" architecture allows it to make highly customized servers in a rapid fashion with extremely quick time to market. Moreover, Super Micro has been an innovator in terms of energy-efficient server designs, making it often the lowest-cost option for data center operators when factoring in electricity and cooling needs. Super Micro's latest feature is liquid cooling systems, which will probably be needed in most AI-oriented data centers going forward due to their massive energy consumption and heat generation.
Management has forecast about 40% growth in the current year, but that seems like just the beginning. Last quarter, Super Micro noted over 50% of its revenue came from artificial intelligence server systems, which is a much higher proportion than rival server makers. That means AI growth will have a much greater impact for Super Micro than rivals.
One last positive factor is that Super Micro is still run by its founder Charles Liang, who owns 14.3% of the company's stock, with a long-term incentive plan that's also highly attractive for current shareholders.
Broadcom
Like the aforementioned stocks, Broadcom (AVGO 3.84%) has also performed well in 2023, And while it trades for a reasonable multiple around 24 times forward estimates, those earnings estimates are probably too low. That's because it's hard to quantify Broadcom's recent $69 billion acquisition of VMware, which just closed in November.
While VMware didn't have much in the way of earnings on its own, Broadcom is already slashing costs and folding the software giant into its corporate structure. Earlier this month, two Wall Street analysts wrote they think the VMware acquisition could contribute $12.50 in earnings per share to Broadcom in the year ahead, which would catapult its earnings much higher than the average analyst estimate today. In fact, these two analysts think Broadcom could earn $60 per share in 2024, which would mean the stock trades at just under 19 times those estimates -- below the multiple of the market.
VMware could be highly valuable in the age of AI, as a software platform that can link different public and private clouds with on-premise data centers in a seamless fashion. Going forward, companies will likely need to store and transfer data cross different platforms to tap into different AI capabilities offered by each, increasing VMware's importance.
In addition, Broadcom's networking segment will also get a big AI boost. Broadcom's leading switching and routing chipsets will be necessary for shooting data around data centers at lightning-fast speeds needed by AI computing. Secondly, Broadcom owns custom AI accelerator intellectual property currently used by several cloud giants that design their own custom AI chips. While Alphabet uses Broadcom's IP in its Tensor processing units (TPUs), other cloud giants are also following suit in making their own custom chipsets, increasing the opportunity.
While CEO Hock Tan noted AI-related revenue accounted for just 15% of Broadcom's semiconductor revenue in 2023, he expects that proportion to grow to over 25% in the year ahead. That makes it a top AI stock to own beyond the Magnificent Seven as well.
