Fears about over-concentration in the US equity market and a potential AI bubble are overblown, Unisuper CIO says
The concentration of the US equity market into a few extremely large technology companies has been weighing heavily on the minds of institutional investors.
Not only has it made active management, and hence the ability to outperform the market, difficult, but it has also increased investment risk dramatically. After all, it only takes a handful of stocks performing badly to drag down the entire equity portfolio.
But in a world where peer-relative performance is dominant, there is little that Australian super funds can do to take preventative action against a technology meltdown, as the old adage of ‘the market can stay irrational longer than you can stay solvent’ rings true (although in a Your Future, Your Super world, super funds might want to replace ‘solvent’ with ‘compliant’).
It means many investors anxiously await any indication of cracks starting to form in the technology industry.
But John Pearce, Chief Investment Officer of the $158 billion Unisuper, remains convinced that technology, including artificial intelligence (AI), will continue to drive growth and that the current dominance of US technology companies is simply a symptom of a long-term, structural trend.
This fear of over-concentration, I just think is misplaced. If you look through history, you'll always have market leadership
“This fear of over-concentration, I just think is misplaced. We know the Mag Seven is about 30 per cent [of the S&P 500 index] right? If you look through history, you’ll always have market leadership,” he says in an interview with [i3] Insights.
Pearce pointed to the dominance of Cisco in the late 1990s and technology, media and telecom (TMT) stocks in the early 2000s, when the internet took off. “Now, you’re seeing another inflection point in that revolution with AI,” he says.
“Why shouldn’t we have a massive leadership by the tech sector? To me, it’s pretty obvious that we should, and I’m not particularly concerned about it. You know, I’d be starting to get concerned if we see valuations getting to the 2000, dotcom level of valuations, but we’re nowhere close to that. If you look at free cash flow yields, then they’re still really healthy.
“The risk, of course, is that free cash flows are starting to get used only on AI, in the chase for AGI (Artificial General Intelligence),” he says.
Show Me the Money
But not everyone is so bullish. A report by MIT, called ‘The GenAI Divide: State of AI in Business 2025’, found that 95 per cent of AI pilot projects yielded no discernible benefits.
Although the MIT report has sometimes been interpreted as saying that 95 per cent of AI companies don’t make any money, the report actually points out that most companies fail to implement AI successfully into their operations, whether they bought an external product or developed one of their own. It shows the end client doesn’t see any benefit.
The key problem, the authors say, is that many AI systems don’t learn (improve) and don’t integrate well into existing processes.
It is not due to a lack of trying. “Enterprises are extremely eager to adopt AI and 90 per cent have seriously explored buying an AI solution,” the authors of the report write.
In fact, the researchers found that every single one of their respondents used some form of AI in their jobs, but mostly in a personal capacity in an effort to reduce their workload, and without their managers necessarily knowing they did so. It is a development the researchers call the “shadow AI economy”.
The few companies that did integrated AI well shared a few common characteristics, including the fact that they bought a system rather than build one, empowered line managers rather than have central labs and selected tools that integrated deeply while adapting over time.
Ultimately, this will give rise to a web of AI agents that can interact with each other seamlessly and self-optimise across platforms. For example, an early experiment showed how procurement agents identified new suppliers and negotiated terms independently.
When does the music stop? I don't see it stopping anytime soon, because you've seen this latest reporting season, you've seen how much money, how much cash, is still being made by these big tech companies, and they're not pulling back
“The most forward-thinking organisations are already experimenting with ‘agentic systems’ that can learn, remember, and act autonomously within defined parameters,” they said.
Yet, most companies are nowhere near this utopian vision of AI implementation.
Pearce acknowledged that most AI pilots fail, while the technology companies that have invested in AI companies haven’t seen an adequate return on their investments either. But that doesn’t mean technology companies are likely to give up any time soon.
“If we were actually charged the price to use Copilot, even a cost recovery, a lot of people would opt out. It’d be so damn expensive,” he says.
“But when does the music stop? I don’t see it stopping anytime soon, because you’ve seen this latest reporting season, you’ve seen how much money, how much cash, is still being made by these big tech companies, and they’re not pulling back.
“There’s clearly a sense of FOMO (fear of missing out). There’s a fear that the first to crack artificial general intelligence will take it all, that the winner takes all. That fear alone is going to keep these big tech companies in the race for at least the next couple of years.
“So I can see the party continuing in the short to medium term, but, yes, if that final demand and pricing doesn’t come through, then a lot of money is going to be lost,” he says.
Is It All Just One Big Ponzi Scheme?
The fear of final demand not coming through is one of the key concerns of many investors, but it is not the only one. There are also fears over the structure of investment into AI companies, in particular the idea that some of these investments seem to be circular in nature.
Fund manager GQG has pointed out that there is a dependence between venture capital (VC) firms that invest in AI companies, the AI companies that pay for access to computing power with cloud companies, and the cloud companies that invest heavily into AI-focused venture capital funds.
“It is … unclear how much of the cloud sector’s revenue growth now comes from AI startups, which are typically funded by the same cloud companies. According to one estimate, AI startups spend more than 80 per cent of raised venture capital money on compute resources,” GQG said.
“If VC funding dries up, a substantial slice of the cloud sector’s incremental spend could be at risk, potentially slowing growth.”
Asked if he was concerned about investments into AI ending up being a house of cards waiting to be brought down, Pearce was dismissive of those who argued the AI boom was akin to a Ponzi scheme.
I've seen the word Ponzi being used to describe it; this is not Ponzi. Ponzi is when you're getting buyers to come in and there's nothing at the other end. It's just fictitious. There are no assets. There's something real going on here [with AI]
“I’ve seen the word Ponzi being used to describe it; this is not Ponzi. Ponzi is when you’re getting buyers to come in and there’s nothing at the other end. It’s just fictitious. There are no assets,” he says.
“There’s something real going on here [with AI],” he says.
Again, Pearce compares the current situation to the internet boom of the early 2000s and points out some questionable metrics that were being used to measure success rather than the tried and tested company fundamentals.
“I remember the valuation metric that was being used during the dotcom [bubble] was ‘eyeballs’. But when you talk about the big tech companies that make up this 30 per cent [of the market] now, then you’re talking about free cash flow yields. That’s a completely different metric than what we were using during dotcom; I can’t recall people talking about free cash flow yields then,” he said.
Who Dares to Go Naked CBA?
Although the Australian equities market is quite different from the US, not just in size but also in sector diversity, it also battles concentration issues. The most obvious one is the dominance of Commonwealth Bank of Australia (CBA). CBA is about 12 per cent of the ASX 200 index, by far the largest stock in the index, and is trading at valuations that make it one of the most expensive banks in the world.
Pearce is well aware of the difficulty balancing the fund’s views on valuation with the demands of a peer aware industry.
“Look, at the end of the day, we got this thing called the performance test, so whether people like to admit it or not, it has made us all a bit more benchmark aware,” he says.
“The industry has probably got a structural underweight to CBA already. Does a fund want to go more underweight or even naked CBA? That’s a hell of a risk. Now bear in mind, CBA is by far the best bank of the four, but on any objective valuation metric it’s ridiculously expensive, so we’re still underweight,” he says.
Pearce is not planning to change his underweight position, but acknowledges it becomes more difficult every day to hold on to the position, let alone increase its underweight to the bank.
“We’re holding our nerve. But are we getting even more underweight? I don’t think we’ve got the nerve to do that.”
Some super funds in Australia have decided that they don’t want to second-guess large caps and have implemented a passive allocation to the top 20 or 50 stocks, while taking a more active approach in small and mid-cap companies.
Unisuper also holds a number of passive positions, especially where it serves to manage excessive risks in the portfolio. But Pearce is reluctant to cancel out any active bets.
“We do take passive positions,” he says. “If we find that our aggregate underweight is way more than our risk tolerance allows, then we could take a passive position, but we certainly wouldn’t be neutralising [any positions],” he says.
_________
[i3] Insights is the official educational bulletin of the Investment Innovation Institute [i3]. It covers major trends and innovations in institutional investing, providing independent and thought-provoking content about pension funds, insurance companies and sovereign wealth funds across the globe.