Artificial intelligence (AI) has been a buzzword over the past six months, but the AI buzz is much more real than the fly-by-night trends like altcoins, meme stocks, and SPACs seen over the past few years.

Transformer technology -- in which an AI can "supervise" its own training without a human labeling everything -- and advances in semiconductor technology have combined to make human-like AI a reality.

The astonishing capabilities of ChatGPT (released in November 2022) appear to have kicked off a new AI arms race. While leaders in the field, such as Nvidia (NVDA -6.62%) and Microsoft (MSFT -1.33%), have seen their stocks surge to high multiples, the following three AI-related names still trade at cheap valuations. That means their stocks could go parabolic as the AI era barrels forward.

Meta Platforms

Meta Platforms (META -5.68%) stock may not look cheap at the moment, but its headline price-to-earnings (PE) ratio of 25 is a bit misleading. 2022 saw a "perfect storm" for Meta, including a global economic and advertising slowdown, headwinds from the Apple iOS privacy changes that hurt Meta's ad targeting capabilities, and the competitive threat of Tik Tok. Oh, and Meta Platforms also lost $13.7 billion last year on its metaverse venture, which doesn't really yield much revenue and has a highly uncertain payoff.

Despite that perfect storm, Meta still made about $29 billion in operating profit last year and $42.7 billion in the Family of Apps (FOA) core segment, excluding metaverse losses. And those figures are down from their 2021 peaks of $46.7 billion in total operating earnings and $56.9 billion in FOA operating earnings. So at Meta's current market cap of $548 billion, the stock still trades at less than 10 times the core business' peak 2021 operating profit.

Keep in mind that Meta was a 20%-plus grower in the years prior to the recent slowdown. If it can return to that type of growth rate, the stock could explode.

How does Meta plan to get that mojo back? In short, artificial intelligence.

Human face emerges from machine-like pixels.

Image source: Getty Images.

On the advertising front, Meta is using AI to help improve its targeting capabilities following the iOS headwinds since late 2021. For instance, a new AI program called Advantage+ is helping advertisers recover their returns on Facebook and Instagram ads. In the program, advertisers allow Advantage+ to rapidly iterate ads by tweaking text or images and then test them at scale before widely distributing the best iteration. According to a recent Financial Times article, Advantage+ has already helped some advertisers regain the returns on ad spend lost to the iOS targeting change.

AI has also helped Meta's Reels product take off. The short-form video format powered by Meta's AI Discovery engine has doubled its engagement over the past year, effectively negating the Tik Tok threat. And on the recent fourth-quarter conference call, CEO Mark Zuckerberg noted how AI is helping Meta cut its bloated cost structure while helping software engineers write code faster and be more productive.

Beyond this year's results, Meta also just released its own internally developed large language model to researchers and recently formed a high-level product group, pulling AI engineers from various parts of the business into a single group set on AI innovation. In early April, Meta released a new AI model that accurately identifies items within images.

So investors should expect Meta to benefit from AI not only in the near term with better engagement, advertising monetization, and lower costs but also in the future as it develops new AI tools for creators and potentially whole new businesses. With its stock still at an undemanding valuation based on its core business profits, Meta's stock could very well go parabolic.

Super Micro Computer

Meta is actually a big customer of server assembler Super Micro Computer (SMCI -6.92%), which provides customized server solutions for large and small enterprises.

While Super Micro has sort of already gone parabolic, with its stock up a whopping 146% since Jan. 1, 2022, it still only trades for less than 10 times earnings!

SMCI Chart

SMCI data by YCharts.

Server companies generally trade at low multiples, but Super Micro's CEO Charles Liang sees big-time growth for the company going forward. Even in the current soft year for server sales, Super Micro should see 30%-plus growth, and Liang sees at least 20% growth in 2024 on top of that.

How has Super Micro pulled this off? Well, first, the company was starting with a low earnings base after an accounting snafu in 2018 caused Super Micro to pause growth initiatives to get its accounting house in order. However, that accounting trouble had to do with the timing and reporting of revenue, not fabricating sales or earnings.

After being reinstated to the Nasdaq in early 2020, Super Micro got back to business developing innovative "rack scale" solutions that cut costs for data center operators, opening a new manufacturing plant in Taiwan to lower costs, and increasing investments in its proprietary "green" computing solutions.

Unlike many server companies that work with standard mass-produced models, Super Micro takes a different approach with its "building block" architecture. That modular design and the company's close relationships with Silicon Valley chipmakers mean Super Micro is often first to market with servers containing the latest chips, such as Nvidia's H100, while also having mass-customization capabilities for customers wanting to design application-specific servers. Moreover, Liang has long heralded a "green" computing ethos, with energy-efficient designs allowing for lower energy consumption and less heat generation.

Combining the ability to customize at scale, lower electricity bills for data center operators, and be first-to-market with new chips caused demand to take off last year. And those advantages should only be magnified in the age of AI. As more large companies look to experiment with AI, Super Micro's ability to lower escalating power bills and provide tailored server solutions should be a massive benefit.

That's how Super Micro has defied the tech slump over the past year. But with the AI wars just beginning to take off, I think Super Micro can continue to march higher, given its still-discounted valuation.

Letters A and I on a sea of green electric currents.

Image source: Getty Images.

Lam Research

Finally, semiconductor equipment manufacturer Lam Research (LRCX -10.07%) is also a bargain-priced stock set to benefit massively from the growth of AI.

Among the semi-cap equipment manufacturers, Lam is distinguished from the pack in several ways. First, it has highly advanced technology for the vertical stacking of semiconductor components. That has helped it gain massive share in NAND flash production, which began vertically stacking modules 10 years ago.

Unfortunately, that has given Lam an outsized exposure to the memory industry, which is currently in one of its worst-ever downturns, causing Lam's revenue to fall 32% quarter over quarter in March, with management projecting another 20% decline in the June quarter.

But it's also why Lam trades at a valuation of just 14 times earnings -- a veritable bargain for a company that should benefit from the long-term growth of AI.

First, AI is memory-intensive, with AI servers requiring several multiples of DRAM and NAND content per server than typical servers sold today. AI servers only make up about 1% of all server shipments but are expected to grow at a double-digit annualized rate over the next five years, which should eventually lead to a memory recovery and, therefore, Lam's memory business.

But leading-edge logic chips for AI are also going vertical now, especially with the new type of transistor architecture called gate-all-around (GAA). Lam believes it will take share in GAA transistor production, which will lead to an incremental $1 billion revenue opportunity.

And Lam's dry resist technologies are increasingly used with extreme ultraviolet lithography (EUVL) necessary for advanced chip patterning. Last quarter, logic and foundry equipment sales comprised two-thirds of Lam's sales, whereas, in prior years, memory garnered that large majority of Lam's revenue.

So a memory recovery, along with increasing share gains in advanced logic chips for AI, will benefit Lam in the future as the market recovers. With a beaten-down valuation and Lam still quite profitable even during this slump, look for the stock to take off as the AI wars kick into high gear.