How to Invest in AI

digitally generated image of multi-colored gear wheels in the shape of a brain

(Image credit: Getty Images)

published 29 July 2023

Artificial intelligence is already writing our kids' essays, driving cars, producing new computer programs and helping scientists seek cures for cancer. It also promises lower costs and bigger profit margins for millions of businesses.

By 2030, a new report from the consulting firm McKinsey predicts AI could increase global productivity by more than 3%, adding trillions to the world's wealth. Microsoft (MSFT) founder Bill Gates, who knows a thing or two about technology, calls AI a "revolutionary" development that will likely soon "be able to do everything that a human brain can, but without any practical limits. This new technology can help people everywhere improve their lives."

Subscribe to Kiplinger’s Personal Finance

Be a smarter, better informed investor.

Save up to 74%

https://cdn.mos.cms.futurecdn.net/hwgJ7osrMtUWhk5koeVme7-200-80.png

Sign up for Kiplinger’s Free E-Newsletters

Profit and prosper with the best of expert advice on investing, taxes, retirement, personal finance and more - straight to your e-mail.

Profit and prosper with the best of expert advice - straight to your e-mail.

Extravagant claims about the potentially world-changing impacts of computers that can learn and reason on their own are raising hopes – and fears – around the world. The hype is also sending the stocks of AI-related companies soaring. The tech-heavy Nasdaq Composite index rose at almost double the S&P 500's 17% pace in the first half of 2023. And most of that was driven by a handful of AI-related firms. Nvidia (NVDA), which designs the chips used in more than 90% of AI applications (according to technology industry research firm CB Insights), returned an eye-popping 190% over the period. Many AI-related stocks trade at more than 40 times next year's expected earnings – about twice the level of the overall market.

That creates challenges for investors interested in benefiting from AI but afraid of a repeat of 2000, when the dot-com bubble burst. Big technological innovations "are often overhyped in the near term," warns Dan Fletcher, portfolio manager of the DWS Science and Technology Fund. But "they are underappreciated for the long term." He predicts that AI will have "very profound effects" over the next decade.

Right now, "the market is indiscriminate," Fletcher says, inflating many companies with only tangential connections to AI, whose transformative promise is only beginning to be understood. But there's still opportunity for disciplined investors who do their homework and either have an investment horizon long enough to withstand potential volatility or who can wait to buy during the inevitable dips. Veteran tech analysts say investors should ask three key questions to find companies with good chances to win the AI race:

Longtime portfolio manager Gregg Fisher, who currently runs Quent Capital, an asset management firm, says he checks the résumés of company leaders and in-house technologists to see if they really have AI expertise. And he looks at company statements from the past several years: Did the company use data analytics and machine learning before AI became all the rage? And, importantly, what percentage of a firm's revenues and profits are AI-related?

Are the business and the stock fundamentally sound? Ryan Jacob, manager of the Jacob Internet Fund, says one lesson from the dot-com bust is that investors who ignore fundamentals will get burned. A strong balance sheet showing that a company has the resources to compete and expand is a must, he says. As for stocks, recognize that fast-growing companies should be analyzed with growth and opportunity in mind, says Fisher. But don't fall for claims that "price-earnings ratios don't matter anymore," he says.

What part of the AI chain does the company serve? Businesses in nearly every sector will be experimenting with ways to profit from AI. Many, if not most, of those efforts will fail. Many experts say that for now, it is safer to invest in the companies that provide AI tools to the experimenters. DWS's Fletcher says that means generally focusing on firms that design or make AI-related equipment or software for others to use. "We think infrastructure is the first place to go," he says.

Chip designer Broadcom (AVGO), a big player in cable set-top boxes and ethernet routers, has ramped up production of AI-related semiconductors. Chief executive Hock Tan told investors in June that AI revenue will likely hit $4 billion in fiscal 2023 (ending in October), making up about 11% of total annual revenue. He expects AI-related sales to soar to as much as $7 billion in fiscal 2024. The firm is on pace to earn more than $42 a share in fiscal 2023, up 12% from the prior year and about 50% from fiscal 2021.

Broadcom has benefited from this year's AI stock boom. The shares jumped 57% in the first half. But in part because Broadcom's profits are booming as well, the stock's price-earnings ratio based on expected earnings was below the tech industry average of about 20. That is one reason the company is a favorite of the Columbia Seligman Technology and Information Fund, whose mission is to find "growth at a reasonable price." The semiconductor stock is one of the fund's top three holdings. Comanager Shekhar Pramanick says Broadcom should benefit in 2024 as firms increase spending on AI to infuse it into more of their processes. "Everybody wants to experiment," he says.

Angelo Zino, senior stock analyst at CFRA Research, expects annual AI-related sales at Microsoft to reach $10 billion by the end of 2023 – equivalent to almost 5% of the company's expected $211 billion in annual sales – then "grow exponentially" for the next few years. Microsoft "is an obvious play" for any investor who wants to jump on the AI train, he says. Its investment in OpenAI gives Microsoft a say in which companies will get access to that system. Combine that with the company's huge size and dominance in areas such as office software, and you get "extreme pricing power," he says.

Dominic Rizzo, manager of T. Rowe Price's Global Technology Fund, likes the prospects for Microsoft's "Copilot" AI services. The applications are already helping programmers generate code faster and are being added to Windows and Microsoft 365 to help office workers with tasks such as creating Excel spreadsheets and PowerPoint presentations. Microsoft makes up more than 10% of the fund's portfolio. Though the shares trade at 32 times expected earnings, Rizzo calls the premium P/E "reasonable when you consider Microsoft's best-in-class business model, customer relationships and AI positioning."

The chip industry overall suffered as sales of cell phones and laptops slumped over the past year. But Nvidia's sales have skyrocketed from $11 billion in the fiscal year ending January 26, 2020, to almost $27 billion for the most recent fiscal year. For the quarter ending April 30, the company reported earnings of $1.09 a share, down from $1.36 a share in the same quarter a year ago but well above the 92 cents a share expected by a consensus of Wall Street analysts. Sales were $7.2 billion for the quarter, compared with expectations of $6.5 billion. Nvidia executives said they expected sales of roughly $11 billion for the July quarter.

Nvidia's potential is no secret – the stock has shot up from roughly $100 a share in the summer of 2020 to $423 recently. It remains well loved on Wall Street, with nearly three-fourths of the analysts who follow the stock recommending it. But the run-up has led many investors not yet holding Nvidia shares to ask whether they've missed out. Others worry about the firm facing increasing competition or stricter regulation here and abroad.

Whether to buy now depends on your time horizon, says Matthew Moberg, manager of the Franklin DynaTech Fund, which has 6% of its portfolio in Nvidia shares. The stock, trading at 47 times expected earnings, will likely be volatile and is not a good choice for short-termers at current levels, he says. "If you have a long-term investing time frame – five to 10 years – you shouldn't worry about the run-up," he says. "I feel really good about Nvidia. It still has room to grow."

Demand for new kinds of AI chips is expected to increase the company's revenues by more than 14%, to $5.8 billion for fiscal 2023, which ends October 31; earnings are expected to jump over 20%, to nearly $11 per share. The stock has returned 36% so far in 2023. Despite the gain and a P/E of 37, Mobley says he thinks the stock is "still relatively undiscovered." Fletcher, the DWS portfolio manager, says Synopsys is in great shape to ride the AI boom, with no debt and very few competitors, which gives the company some pricing power.

For now, the shares attract value-seekers such as T. Rowe Price's Rizzo, who has made the company the fund's third-largest holding. "Taiwan Semiconductor is the best-in-class semiconductor foundry," he says, adding that AI will power significant growth for the company over the next 18 to 24 months. Morningstar analyst Phelix Lee agrees that the shares are attractively priced, trading about 30% below his estimate of their fair value.

Kim Clark is a veteran financial journalist who has worked at Fortune, U.S News & World Report and Money magazines. She was part of a team that won a Gerald Loeb award for coverage of elder finances, and she won the Education Writers Association's top magazine investigative prize for exposing insurance agents who used false claims about college financial aid to sell policies. As a Kiplinger Fellow at Ohio State University, she studied delivery of digital news and information. Most recently, she worked as a deputy director of the Education Writers Association, leading the training of higher education journalists around the country. She is also a prize-winning gardener, and in her spare time, picks up litter.