Morgan Stanley Investment Management’s Dennis Lynch (main picture) is not betting on a clear winner for artificial intelligence (AI) just yet.
“We’re not sure where there’s a clear investment case for a company long term at this point in time,” Lynch said in an interview. “I think there are questions about the end-market demand.”
The strategies that Lynch runs have delivered returns that rank in the top 1% of funds in their category in Hong Kong and Singapore, according to data from FE fundinfo.
But unlike most top-ranked growth equity funds, they have managed to outperform without overweight positions in AI market darlings such as Nvidia, Meta and Microsoft.
While AI technology and large language models (LLMs) are impressive, Lynch is sceptical on there being a clear AI ‘winner’, pointing to “the absence of very profitable products” that can support the necessary capital expenditure on the latest chips.
Nvidia’s products however, are very profitable, and its surging sales have made the company by far one of the biggest beneficiaries of the race to build out AI applications.
But Lynch said there are questions about the sustainability of this chip demand: “Our big conundrum has been trying to understand whether it’s a one-time supportive infrastructure build, or is it a continuous build, in which case it’s more secular as opposed to cyclical – that can be really hard to determine.”
This is why his strategies haven’t made any investments based off the progress of LLMs or any of the infrastructure companies that support the training and execution of those models.
He said his team is “not sure yet that there’s a clear, unique company that has a sustainable competitive advantage”.
“In my view, there’s a general absence of end user products that are differentiated,” he said. “There’s tremendous benefits [of LLMs], but also some mitigating factors that make it hard to build concrete products that lead to revenues and cash flows.”
Although LLMs have been widely adopted, Lynch noted that consumer interest in using LLMs has started to fade over the past twelve months.
Some of the most widely used LLMs such as ChatGPT, Microsoft Copilot, Amazon’s Claude and others are struggling with rising hallucination rates, where the chatbot generates fabricated and false responses.
Lynch said: “On the corporate side, people in corporations want things that work. There can be a little bit of an error rate, but it has to almost be predictable.”
“With LLMs there’s a lot of unpredictability when the errors kick in, and that makes it hard to incorporate it into a more commercial environment.”
Where Lynch sees opportunities
While many top-performing funds have been touting their bets on the mega-cap technology stocks sometimes described as obvious AI beneficiaries, Lynch thinks the best opportunities are in the less widely owned stocks.
“My general sense right now is that the younger companies that are less profitable in the short term, but have big end-game potential, still offer a really strong upside,” he said.
Due to a preference for smaller high growth stocks, Lynch’s growth strategies experienced a torrid year in 2022, but they’ve since bounced back to deliver top-ranked returns in 2023, 2024 and year-to-date.
“2022 was hard to live through,” Lynch said. “We’ve stuck with the companies coming out of it, and there’s evidence that that was the right thing to do so far.”
“Our success since then still doesn’t match up with where we want to be, but I think it’s all about when you’re in a spot, making the right incremental choice, which for us is continuous ownership.”
Lynch said he likes to stick with people, companies, and relationships for long periods of time, admitting that it’s probably “a bias” of his.
“I’ve been dating my wife for 35 years, married for 29, and have supported the same football team my whole life: the Pittsburgh Steelers.”
He said: “We like to make continuous long investments in companies that are unique, but we also like to do that with our people and we do our best to keep them around. If you look at our team, you’ll see that there’s a lot of continuity when it comes to people.”
Where Lynch does see opportunity is in what his team call tailwinds. One tailwind he likes in particular, is “distributed technology networks”, which is offered as a service to promote cloud computing.
“We have an idea in the portfolio we think can really benefit from the way the world’s evolving, in the sense that it wants more networks of servers opposed to direct servers,” he said.
Another area which isn’t necessarily groundbreaking, but Lynch suggests is perhaps still underappreciated, is e-commerce.
“E-commerce continues to be an area that I think local scale and/or product specific scale can lead to unique long term competitive advantages,” he explained.
“Sometimes things correlate for a period of time, and then they don’t.”
Despite stellar performance, holding Lynch’s strategies are not for the faint of heart, as the outsized returns have come hand in hand with higher market volatility.
But the recent wild swings in equity market volatility has been made worse by the rise of factor investing, according to Lynch, who is critical of the effect that factor-driven trading strategies have on the market.
Factor investing strategies use an approach that picks stocks based on “factors” such as value, momentum, size, quality, or volatility.
“Factors are generalising, not thinking of companies as idiosyncratic or unique, but thinking of them as themes or generalising about their nature,” he said.
“That’s not investing in my mind, it’s more an abstraction of investing, it’s layer two. Whereas company knowledge, company specific fundamental research, is layer one.”
“In times when there is high market volatility, I think a lot of short-term market participants who tend to rely on those layer two abstraction approaches treat everything with a broad brush in a very reactive manner.”
In Lynch’s view, market prices should be driven more by fundamental-driven decision making, rather than factors.
“I think it’s fine when it’s the minority of decision making, but whenever the tail starts wagging the dog, you do get some weird stuff.”
“Most factor analysis is more about prices, how prices correlate with each other for certain stocks and sectors. It reduces companies to something that’s more about price-driven thinking.”
“The problem with that is sometimes things correlate for a period of time, and then they don’t. It’s sort of assuming that the correlation between two things make them equivalent, and the problem in markets over time is when people get too confident in these assumptions.”
He said that sometimes investors run the risk of betting on a connection between two groups of assets that may suddenly no longer correlate – “because maybe the fundamental connection wasn’t there to the degree that the prices suggested”.
Quality growth investing has become formulaic
One of the most popular investment approaches of the past decade has been to buy and hold so-called “quality” companies with certain financial fundamentals like high return on invested capital (ROIC), growing earnings per share and consistent free cash flow.
But Lynch suggested that this style of investing has become so crowded to the point where most of the potential outperformance has been arbitraged away.
He said: “During my time at Columbia Business School, they really preached the value of return on invested capital. At the time, the message was if an investor simply bought stocks with a hurdle rate of 10% ROIC and 10% growth, you would win versus the benchmark.”
“Here’s the problem: when it’s that simple, and it was for maybe a window, what happens is everyone starts figuring it out. I would argue, in today’s world, everyone’s over-excited about those companies that tend to be called quality and that tend to be low volatility.”
“What can go wrong there is that investors can miss something larger that’s happening, particularly when you’re looking at older companies that have a long history, where people feel like it’s a durable situation or durable franchise.”
He said because the growth rates of these type of companies aren’t usually that high, buying at the right price becomes critical. But in focusing on price, investors risk missing a fundamental change that might be occurring with the business.
“Sometimes what happens is people get excited about the price in those cases but what’s changing is so big, the company is not equipped to deal with it,” he said.
“A lot of people present their process almost like an algorithm. They show a funnel, and they show certain financial characteristics they care about, and I think it’s sometimes overfitting something that’s not quite that specific.”
“It’s not really the process, it’s more the people, the philosophy, and the execution that matter. It’s easy to talk about being long term. It’s hard to keep consistently doing that in the face of volatility.”
He believes investors are overly focused on low volatility at all costs, which creates a tremendous opportunity for people that are willing to accept some volatility.
“Drawdowns stink, but sometimes resilience can be built from them,” he said. “All big companies that wind up being great, usually have that happen to them.”
“If you’re thinking beyond a few years, then you actually want some volatility, because volatility is energy and opportunity, and it can be a positive thing.”
