Despite a rough couple of months for stocks that are seen as a play on the artificial intelligence boom, Morningstar analysts see strong demand and monetization of AI functionality as supporting strong long-term prospects for the industry.
There's also good news for investors looking to put new money to work in AI stocks. Amid the pullback in AI stocks, some of the markets biggest names are now trading in undervalued territory according to Morningstar analysts. (A list of undervalued AI stocks can be found at the end of this article.)
The AI Stock Retreat
After leading the market higher during the second quarter, the largest AI names have been dragging the market down in the third quarter, causing many investors to rotate out of big tech stocks and into value stocks.
Names such as Nvidia NVDA and Microsoft MSFT are all down over the past two months, despite beating their FactSet consensus estimates for second-quarter earnings and showing no signs of slowing down on AI progress.
The Morningstar Global Next Generation Artificial Intelligence Index plummeted 14.9% from the July 16 market high to the most recent market low on Aug. 5, nearly double the 7.8% loss of the broader Morningstar US Market Index.
Despite making up some of the losses since then, AI stocks have been a drag on market returns over the past two months. Since July 16, the 12 largest detractors from the return on the Morningstar US Market Index are all technology stocks with strong AI connections.
AI Demand
Investors should not confuse the recent underperformance of AI stocks with evidence that AI-related demand is slowing, according to Morningstar analysts. For example, on the hardware side, "orders are still strong for Nvidia and it feels like demand remains strong through at least 2025, even with Blackwell production delays," says Morningstar director of equity research Eric Compton.
"Hyperscalers are still talking about spending materially in 2025," he adds, which supports the claim for continued demand for chips.
Additionally, OpenAI's recent release of its "Strawberry" model, o1, should help fuel demand even further, according to equity strategist Brian Colello.
"This new o1 model is expected to better handle complex questions and reasoning, with more of a focus on providing thoughtful answers, rather than quick answers," he says.
"From the chip side of things, these types of Strawberry models should require more inference computing power, which continues to support the exponential growth we still anticipate for Nvidia and AMD AMD in AI accelerator chipsets."
AI networking companies, such as Broadcom AVGO, Arista Networks ANET, and Marvell MRVL, also continue to benefit from strong demand, says equity analyst William Kerwin. "We see this growth being durable into the medium-term, and strong through 2028 across these supporting infrastructure players," Kerwin adds.
Early Signs of Monetization Progress
On the software side, "we are starting to see some more instances of software companies able to use/monetize AI functionality, whereas previously we had more doubts about this," says Compton.
One example is at Apple AAPL, where Apple Intelligence, a suite of AI features built into iPhones, is set to be released in early 2025. "Apple is building its own models and partnering with ChatGPT to drive AI features," Kerwin says. "Only the latest iPhones will be able to run these AI features, so we expect a strong growth cycle in 2025 for Apple as consumers upgrade to use AI."
Microsoft and Amazon.com AMZN are two more examples of software companies monetizing AI, specifically through their public cloud offerings, says senior equity analyst Dan Romanoff. However, he notes that despite lots of product announcements and lots of talk around generative AI interest from clients, "from a pure software perspective, only a few companies are enjoying a slight uptick in revenue from AI-related demand."
Valuation
Despite the bumpy road for AI stocks over the past few months, the Morningstar Global Next Generation Artificial Intelligence Index is now roughly fairly valued, hovering at a 1% premium to its fair value estimate on a cap-weighted basis.
Drilling down into the individual holdings of the index, 18% of the stocks have Morningstar Ratings of 1 or 2 stars, meaning they are overvalued according to Morningstar analysts. Fifty-four percent are fairly valued with 3-star ratings, and 28% have ratings of 4 or 5 stars, which means they are considered undervalued.
Going back six months, valuations looked much different. In March, 32% of AI stocks were overvalued compared with just 16% undervalued. Then, after the sharp selloff in late July, 20% of AI stocks were overvalued compared with 31% undervalued.
While the overall AI index looks fairly valued, investors can still find individual stocks trading at discounts, Morningstar analysts say. Taiwan Semiconductor Manufacturing 2330, which trades on the Taiwan Stock Exchange, Baidu 09888, which trades on the Stock Exchange of Hong Kong, and Naver 035420, which trades on the Korea Exchange, are all rated 5 stars. Microsoft, Alphabet GOOGL, and Adobe ADBE are among the stocks rated 4 stars. A full list of undervalued AI stocks can be found below.
The author or authors do not own shares in any securities mentioned in this article. Find out about Morningstar's editorial policies.
