How to invest in the right AI stocks

As is often the case, the emergence of a hot technology has investors scrambling to figure out how they can get in on the action.

Artificial intelligence is back in the headlines after OpenAI released ChatGPT for public testing, giving people the first real taste of a technology that could transform the world.

Yet, despite the dramatic headlines, it may be too soon to start viewing AI as the next growth catalyst for companies. For the typical investor, it means that putting money to work directly into AI stocks isn’t as simple as it might seem.

Popular AI Stocks

There’s no question that major tech firms have for years been bulking up their investments in AI. Well-known AI champions include Apple AAPL, Microsoft MSFT, and Google’s parent company, Alphabet GOOGL. Both Adobe ADBE and Salesforce.com CRM have integrated AI into products, and some already use it to remain competitive, such as those in the cybersecurity space like Fortinet FTNT and CrowdStrike CRWD.

The issue for investors, Morningstar analysts say, is that in most of these cases, AI has yet to fully surface as a major new revenue opportunity. They point to only one concrete area in the market today where there’s money to be made from AI, and that’s in semiconductor companies such as Nvidia NVDA.

Still, the AI arms race may only be getting started, say Morningstar analysts. While offerings like ChatGPT have certainly captured people’s attention, it’s not yet life-changing. “We all keep waiting for really advanced AI that is going to blow our minds, and I think we’re still on the cusp of that,” says Dan Romanoff, senior equity analyst at Morningstar.

Romanoff says that while products like Bard AI and ChatGPT are a more advanced version of AI than the general public is used to, it’s not too different from the basic forms of AI consumers have been able to interact with over the past decade, such as Siri or Alexa, or even just basic search engines.

“I think we’re just warming up and maybe entering the first inning now,” when it comes to AI, says Romanoff.

Companies are Investing in AI to Enhance Products

Even as companies look to adopt AI, it may not translate into improvements in revenue or earnings just yet. “I don’t know that there’s a big revenue opportunity immediately available for anyone,” Romanoff says.

“AI is being used as a feature to make existing products more compelling,” he says. However, “It’s still not a silver bullet. You have to first offer a compelling product. Then if you layer in features over time, such as powerful AI, yes it will make the solution more attractive.”

The focus of most public companies’ foray into AI is to enhance existing products and operations. Alphabet and Microsoft are front and center in that respect, with both companies integrating AI technologies into their search engines in an effort to improve results and engagement with users.

Amazon.com AMZN has already incorporated a number of AI-powered features into its Amazon Web Services business, from helping clients in creating chatbots to automating data analyses.

Apple has been one of the leading acquisitors of private AI companies over the past three years, according to a report from PitchBook Data. Those investments could augment several of their products and services, such as Siri, facial recognition, and possibly augmented/virtual reality, says Morningstar strategist Abhinav Davuluri.

Other notable examples of companies investing in AI to enhance products include Adobe with Adobe Sensei, an AI platform that helps product users with projects such as marketing, design, and more. These services are offered as a part of existing Adobe software products and don’t provide an additional revenue stream to the firm.

Likewise, Salesforce has its own AI platform called Salesforce Einstein. While some features may come at additional costs, many of Einstein’s features, such as predictive search—which helps salesforce users personalize product results for their customers—come as a part of some Salesforce solutions.

Still, the incorporation of AI in business operations and products could potentially provide benefits such as increased efficiency and aid in software companies’ abilities to capture market share with more-competitive products and services. That especially holds true for most cybersecurity software companies such as CrowdStrike, Fortinet, and Palo Alto Networks PANW, which all make use of AI to improve their products and drive sales, says Morningstar equity analyst Malik Ahmed Khan.

Finding AI Companies to Invest In

For investors looking to find stocks that tie in more directly to the AI arms race, they may want to heed the adage of investing in the “picks and shovels” suppliers rather than the end corporate users.

The proliferation of AI requires advanced computing resources and is one of the long-term tailwinds that Morningstar’s Davuluri sees most clearly benefiting semiconductor stocks.

Nvidia is at the top of the list of AI beneficiaries, he says. The firm is the leader in providing hardware for advanced computing needs that serve as a crucial component in developing the infrastructure required to power AI: graphics cards, also known as graphics processing units.

Take ChatGPT, which runs off a language model that analyzes “hundreds of thousands of parameters. All of that is being trained on an Nvidia GPU,” Davuluri says “Nvidia’s latest Hopper GPU family that just launched has a lot of custom parts that are optimized to run these language models more optimally and efficiently.” In an investor presentation, Nvidia has outlined that it sees a $150 billion opportunity in AI.

Nvidia competitors Advanced Micro Devices AMD and Intel INTC are also key players in this market. However, while they “sell chips that sit alongside Nvidia GPUs, they’re not necessarily the focal point of the AI workload,” Davuluri says.

While AMD has a history of competition with Nvidia gaming graphics cards for nearly two decades, the firm only recently started producing graphics cards for data center and AI applications in 2020. Intel is a newcomer altogether in the department.

AI is already a tailwind that Davuluri has baked into his valuations for the companies. While the popularity of chatbots has placed AI in the spotlight in recent weeks, it hasn’t had any material impact on the outlooks for those companies.

Among the three, only AMD’s and Intel’s stocks are trading at prices that Davuluri views as undervalued, with AMD trading at a 30% discount to its fair value estimate and Intel at 19%. Nvidia stock is fairly valued, trading at a 10% premium to its fair value estimate.

Another stock in the semiconductor industry that is among Davuluri’s top picks is equipment supplier ASML ASML. While the firm’s lithography equipment is not directly used in AI, the technology is critical in the manufacture of the advanced semiconductor chips used in AI. ASML trades at a 14% discount.