Money managers stand firm on AI stocks in Asia

Artificial intelligence stocks led the market rout that took place in late July and early August, but fund managers say there still are AI companies with structural growth potential, particularly in Asia.

Some investors might associate AI stocks with big names such as Nvidia, Meta Platforms, Amazon.com and Tesla, but the AI universe also includes a slew of manufacturers that enable the technology, the managers said.

“People bucket AI into one big thing, but we look at it as a part of it being cyclical, where shorter-term supply-demand dynamics impact some of our holdings particularly in memory chips because there is still a demand and supply (effect)," said Singapore-based Steve Sun, executive director and senior portfolio specialist at Morgan Stanley Investment Management’s emerging markets equity team. "There is also the structural part of it, such as fabrication, servers and power supply."

MSIM had $1.5 trillion in assets under management as of March 31.

“We do think there are still cases to be made for structural growth in the long term, but I think with this recent pullback… companies are now reassessing what is the return on invested capital,” he said. “If you look at the largest U.S. tech stocks for that matter, they have massive capex plans, and they have to get return on those investments.”

“There’s going to be some questions. We’ve looked through the numbers… At the low-end of spending, (companies set aside) a few billion dollars a quarter for AI spending — however they define it. And there could be as much as $13-$14 billion (in spending) per quarter,” he added.

The tech-heavy Nasdaq saw a series of steep declines in late July and early August, closing at a three-month low on Aug. 7. By then, the Nasdaq had fallen 13.1% since its record high on July 10.

Between July 16 and Aug. 5, the Morningstar U.S. Market Index fell 8.6%, with the Magnificent Seven - Microsoft, Alphabet, Amazon, Tesla, Meta and Nvidia - making up nearly half of that loss, according to a Morningstar report. AI stocks are loosely defined, but the Magnificent Seven are largely considered AI stocks as they have each invested billions into AI research and technology.

Disappointing tech company earnings and a weak U.S. jobs report spurred the sell-off, as investors anticipated hefty interest rate cuts and the possibility of a U.S. recession.

The stock market rout extended to Asian equities as well, as the MSCI Asia Pacific Index fell more than 10% from July 11 to Aug. 5, Bloomberg reported.

But Asia is where AI-related investment opportunities lie, said Mark Williams, London-based lead portfolio manager at Polen Capital's Asia ex-Japan and emerging markets dividend growth strategies.

“It's in the providers of the (AI) infrastructure. We hear a lot of talk about Nvidia, but everything that Nvidia does requires Taiwanese companies to execute,” Williams said.

This applies not only to Nvidia but all other advanced chip makers, he added. They require memory cells, processing units, connectors, reliable power providers, and liquid cooling systems, to name a few — a huge number of them are in Asia, and particularly in Taiwan, he said.

Some better-known AI companies include Taiwan Semiconductor Manufacturing Co., the world’s leading semiconductor maker; SK Hynix, a memory chip maker in South Korea that said in June it planned to invest 103 trillion won ($75.2 billion) in artificial intelligence and semiconductor related chips; Samsung Electronics, the South Korean corporation that has been expanding its chip-making capabilities; Wiwynn, a Taiwanese company that provides data center and cooling solutions; and Lotes, another Taiwanese firm that designs and manufactures components such as connectors necessary for AI machinery.

“What’s happening is you will get challengers to Nvidia either as more affordable, possible equivalents or chips that are as good in terms of processing power, but those also will feed back into similar areas (like) the equipment manufacturers,” he said.

As large companies like the Magnificent Seven continue investing in AI to increase capacity and run large language models faster, bottlenecks in the supply chain will begin to appear.

“Within some of the (printed circuit boards), you’re already seeing that; in copper clad laminates, you’re seeing that. So it’s these niche areas that are the most interesting because you’ve got the potential for multi-year demand companies where I’d say they’re trading at fair value,” Williams said.

Even though Taiwan currently dominates the AI supply chain manufacturing industry, countries around the world have started to develop their own capabilities.

South Korea in April announced that it would invest about $7 billion in AI over the next three years. The U.S. and Japan have also given billions of dollars in subsidies to the Taiwan Semiconductor Manufacturing Company to set up factories in those countries.

China has also made plans to develop self-reliance in the chips industry. For instance, it launched a China Integrated Circuit Industry Investment Fund, backed by six Chinese banks that committed $15.7 billion, which made up 33.14% of the fund’s total capital.

However, these countries will likely face challenges, particularly in finding specialist talent.

“I was in Taiwan about a month or so ago, and engineers are the element that is in demand… And with the current burst that we’ve seen in demand, that’s the area where people are struggling to find the individuals who are qualified enough. And equally, you’ve had decades building a hub in Taiwan, and it’s very hard to replicate that,” Williams said.

That said, countries like South Korea and China are making inroads in certain areas of the AI space, said Elaine Tse, San Francisco-based portfolio manager for the total emerging markets team at Allspring Global Investments, which had $571 billion in assets under advisement as of June 30.

In terms of memory chip companies, the South Korean firms SK Hynix and Samsung are key, she said. Samsung has also started to gain ground in the foundry space, she added.

TSMC is a leader in the creation of application-specific integrated circuits, but all three elements — memory, foundry and ASICs — are necessary for AI to be successful, Tse said.

AI also requires a large amount of power. By 2030, AI-related datacenters will consume 8-10% of total power demand, up from today’s 2-3%.

High growth in power consumption and the need for grid reliability will drive up demand for copper, which we get from Latin America and, to a small extent, China. Power-integrated circuits and cooling systems will also be in high demand, she said.

China also provides grid equipment, power generators and other power-related equipment, she added.

Tse also acknowledged the global sell-off has extended to Taiwanese and other emerging markets stocks, but “if anything, this gives investors the opportunity to be entering at a much more reasonable valuation,” she said.

“The Mag Seven trade at a substantial premium to the EM hardware companies,” she added. The emerging markets-based AI firms typically trade at 15-20 times earnings, which is a discount to their U.S. peers, which are trading at 25-40 times earnings.

“With the pullback, they’re back to the low-teen levels, and we think there’s definite visibility for 20-25% earnings growth for the next few years. So if anything, this pullback has presented very attractive opportunities to investors,” Tse said.

MSIM’s Sun agreed that some of the companies that are more exposed to market cycles may have changed in terms of attractiveness to investors but for longer-term investors, these AI names remain a long-term play.

“At the end of the day, clients want to figure out do I buy or sell leading global chip manufacturers. At this stage, I think for us, depending on your timeframe for one, we still like these ideas. We still think there aren't that many competitors out there, particularly on the fabrication side,” he said.

Even before the pullback, some of these companies were not necessarily overvalued, London-based Seema Shah, managing director and chief global strategist at Principal Asset Management, said while on a trip to Singapore.

Principal AM had $554.4 billion in AUM as of March 31.

Sectors that have indirect exposure to AI, such as data centers, also are an opportunity for investors. At the same time, there are sectors that stand to benefit from the usage of AI, such as healthcare, biotech and agriculture, she added.

“One of the things that we look at is on food sustainability. The AI side of it – combining it with food sustainability concerns is a pretty proposition. Those are your obvious second tier companies or tech sectors that stand to benefit from AI… that haven't yet been appreciated,” she said.