How to invest amid the “shifting sands” of mega-cap tech

In the last 18 months, the financial landscape has been inundated with views about technology more broadly – and Artificial Intelligence, more specifically. Financial media speculation on AI stocks has become something of a sport.

Just last week, we heard that Nvidia’s (NASDAQ: NVDA) share price growth had seen it become the world’s next US$3 trillion market cap company and the world’s most valuable – albeit momentarily. Then, a couple of days later, many of the same news outlets wrote of a “bloodbath” as some US$600 billion was wiped off Nvidia’s market cap overnight.

Amid all the noise, it’s surprising – and refreshing – to hear new views on the topic of technology, which is precisely what I got when I sat down recently with Janus Henderson Investors’ technology investment guru, Denny Fish. He walked me through his team's investment research process in selecting companies for the portfolio.

Fish addressed questions about the concentration risk of investing in AI and his team’s valuation outlook for some of the biggest names. He also explored some of the potential negative scenarios, such as a sudden drop in demand for AI services, and what it could mean for investors.

How companies move from investment ideas to holdings

Each company undergoes a rigorous vetting process that typically takes one to two months before being included in the portfolio. This starts with an analyst building a detailed model that includes cash flow analysis and revenue. The analyst also undertakes multiple conversations with company management and third-party companies.

“Then, sharing that data with the team, having the team poke holes in the idea. Oftentimes, the analyst then goes back and focuses on three or four things that come out of that, in what I call a marinating process,” Fish says.

This process is all part of building confidence in the durability of individual businesses, which in turn intersects with a valuation “that's powerful for us and based on our expectations of what the company's going to do over the next three to five years.”

Sometimes, the company’s current valuation doesn’t align with what Janus Henderson believes it’s worth, or other questions remain. In these cases, further research is conducted, or the idea gets shelved until its share price presents “a more appropriate entry point.

“It's a very disciplined process. It allows us to both populate portfolio, but create a very, very deep bench of ideas that have a full body of work associated with them that could potentially come into the portfolio in the future,” Fish says.

Why concentration isn't always a bad thing for portfolios

With some 38% of the fund in technology and just three non-tech names among the top 10 holdings, some might regard this as a concentration risk.

Fish explains this is simply a result of his team’s analysis: “The fund has become more concentrated over time because technology has come up in the index, but also, we've seen a higher degree of concentration in terms of economic profits generated within the tech sector too, which is, so what's happened.”

This has led to more tech concentration within the fund, and in turn, more concentration among the individual names held from this sector.

“If you look at the financial performance of these companies over time, particularly over the last three or four years, the concentration has been justified because the financial performance has been so strong relative to the rest of the market,” Fish says.

He also emphasises we’re currently in a period of “very significant and profound technology change, and it is incredibly capital- and knowledge-intensive to be able to produce what needs to be produced to advance these cognitive models.”

“I was around during the dotcom bubble and the difference today versus then [is that] the anointed winners, the Microsoft's, Qualcomm, Ciscos, Oracles – they all traded to a hundred times earnings, and their business models weren't nearly as durable as they are today,” Fish says.

“They were book and ship models. They weren't recurring revenue models. Today, despite the stock performance, all of these big companies trade at quite reasonable multiples given the growth rates.”

He emphasises that the average forward earnings multiple for tech companies is now somewhere between 20 and 35 times.

“That is not particularly expensive given the growth that we've seen. And it's nowhere near what you would consider to be bubble territory,” says Fish.

He also points out that while share prices for some of the big tech firms have risen 700% or more, earnings estimates have risen even further.

The biggest topic of internal debate

The biggest issue of debate inside Fish's team revolves around how AI is affecting various other sectors: “Which companies are positioned best to take advantage of AI and which companies are positioned poorly to actually be disrupted.”

Fish says his team divides companies into those on either the right or wrong side of the AI trend.

“We’re trying to make sure we're comfortable with the strategies that the companies on the right side [of AI] are employing to actually take advantage of this tailwind,” he says.

“From there, the big debates are ‘let's be conscious of what's going on. The sands are going to shift among the mega-cap companies'. There will be some winners, there will be some losers. It's not going to be a rising tide lifts all boats.”

Another key area of debate is whether there’s a looming period of disillusionment, once revenues catch up with the capital spending required to develop AI-based services.

He also notes other interesting technology developments that aren’t AI-enabled can “throw fuel on the fire.” For example, the nationalisation or “re-shoring” of semiconductor production among most developed markets.

“That was well underway before AI hit, and now AI has fuelled fire on the capacity that every country thinks they're going to need, to protect their national interest from a defence standpoint, from an economic standpoint and from a social standpoint. We talk a lot about that,” Fish says.

The Janus Henderson team is also constantly focused on finding other companies with unique data sets that might be targeted by AI’s large-language models to help create “new and unique customer offerings.”

The biggest risks to bullish AI forecasts

“Everyone is in an arms race right now…because they understand the impact that AI is going to have on their business, their competitors' businesses, and the opportunity and the threats ahead,” Fish says.

“If you see your competitor spending, you feel like you have to spend too. Right now, all these companies…understand the ROI that they're going to get and they're modelling out their businesses.”

A big question here is whether we could see a scenario where capex is ratcheted up to extreme levels but the revenue from those services comes through far more slowly than is hoped. For example, customer uptake or business roll-out could occur much slower than is currently projected.

“That’s not good for the picks and the axes – that is, the semiconductor makers. It's also not good because all this capex is going in the ground…when investors are expecting strong revenue growth over the next several years,” Fish says.

“The worst-case outcome would be the revenue growth is much slower, that leads to lower capex [and] lower numbers of units procured throughout the semiconductor supply chain.”

In this scenario, declining unit volumes and falling prices would compound.

“I think that would be the scenario where the bears…could run a victory lap for 12 or 18 months, but miss the bigger picture of what's going to happen over the next 10 or 15 years,” Fish says.

Denny's take on the biggest AI beneficiaries

Across Microsoft (NASDAQ: MSFT), Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) and Meta Platforms (NASDAQ: META), Fish notes they’re spending “a tonne of money, roughly $200 billion between the four of them this year."

“The expectation is that starts leading to an uplift in revenue growth in 2025 and beyond, and margins that aren't dilutive to company margins. That's our best-case scenario. We get continued high returns, we get accelerated growth and all is good.”

Turning to Nvidia, Fish notes that many have a “pretty hard fade” in its growth rate, with its most recent growth rates unable to be sustained over the longer term.

“The question is, what's the degree of that fade? And that's why you can get a really wide range of estimates for Nvidia as you start looking out to say, 2026, because it's a very different earnings stream if the fade is 30% versus 10%. That is what investors are really grappling with right now,” Fish says.

“We all know the rest ‘24 is going to be great, ‘25 should be really good. What happens when we start getting through '25 and into '26?”

Fish believes the most important question to consider is whether Nvidia can control the fade of its earnings growth more gradually than many investors expect over the next couple of years. 

Managed Fund

Janus Henderson Global Research Growth Fund

Global Shares

View