Over the past year, the market has rewarded earnings acceleration and surprise, and punished any sign of doubt or prudent hesitancy from company management teams. In short, humility is out. Moonshots are in.
AI is at the centre of that, particularly semiconductors and the companies building the infrastructure around them. That has helped emerging markets and Japan, which are heavily tied to electronics and the chipmaking ecosystem. We've never owned stocks in these regions, as we don’t think we can add value there.
Similarly, oil and gas businesses have done well, both in the third quarter and over the year so far, driven by the energy crisis stemming from the effective closure of the Persian Gulf.
We don't like commodity companies as they take whatever price they're given and can't control their own destiny. Again, we won’t adjust our process because of short-term movements.
We do own plenty of AI champions, including chip designers Nvidia and Arm, cabling and sensors supplier Amphenol and cybersecurity specialist CrowdStrike, but not at anything like the weights they make up in the FTSE World index of global stocks. Some experts reckon AI-linked companies now make up half the US market – we want more diversification than that.
We don't think the answer to a concentrated market is to become concentrated yourself. Sometimes giving up returns is the right call if you don't believe they'll last.
While industrials and consumer discretionary have performed poorly of late, it’s precisely in these areas that we have been finding some of our most attractive investments. When prices fall, valuations improve, and we believe it makes sense to buy when these prices are keener.
CrowdStrike was our standout in the third quarter, its performance a sharp improvement from a pedestrian couple of years.
We bought CrowdStrike in February, when the market had decided all software would be an AI casualty. AI isn't making cybersecurity less important. It's making the threat more sophisticated. Cyberattacks have become faster, cheaper, more deceptive and easier to scale.
Meanwhile, companies are cutting the huge number of security products they use down to a few platforms. CrowdStrike had been caught up in a sell-off far broader than its own fundamentals justified.
Arguably nothing changed with this stock, just the market’s view on how AI will affect it. This is an example of why we think it often pays dividends to ignore the crowd when all the building blocks of success appear to be there.
Changing our minds
Rollins is a good example of where the facts have changed. We thought pest control was a wonderfully resilient business, because people still need it whatever the economy is doing. But AI is starting to change how customers find these businesses through search, which could change the economics of winning new customers.
Management couldn't give us enough confidence about how they'd deal with it, so we sold. Rollins hasn't suddenly become a bad company. The question was whether it was still as predictable as we thought, and the answer had changed.
AI remains a really important theme for us, but it isn't one trade. There are the obvious enablers. Then there are the companies adopting AI, the companies monetising it and the companies protecting it, like cybersecurity businesses.
We've also been adding to what I'd call resource independence. The world is much more conscious of where its energy and critical materials come from and how resilient its infrastructure really is. Fixing that will take a lot of investment.
We’ve bought and added to positions in Caterpillar, mining equipment maker Sandvik, and electrical engineers Siemens Energy and Quanta Services. They give us exposure to infrastructure, electrification, grid investment and industrial spending without needing to guess what copper or oil will cost next year.
We've kept our defensive side too, adding cosmetics company L'Oréal and energy drink business Monster alongside drinks supplier Coca-Cola and price-conscious retailers Walmart and Costco. The growth of these businesses depends far less on everything going right in the economy.
Many consumer stocks look undervalued to me. Rising rates, high gas prices and the cost-of-living squeeze have hit them hard, but those conditions won't last forever. Tougher earnings this year will make next year’s earnings easier to beat. Also, any sustained downturn should hurt their higher-priced and smaller-scale rivals more, leaving them with less competition when the economy comes out the other side.
Take Walmart, one of the most advanced adopters of AI I've seen in our portfolio. More than half its app customers now use Sparky, its AI shopping assistant, instead of typing a search. Ask Sparky for taco night for a family of four on $3 a head, and it fills your basket. Behind the scenes, it runs an advertising auction among the salsa and guacamole suppliers. That's a high-margin advertising event, and Sparky is driving a 40% increase in basket size. It’s the evolution of AI from the enablers to the adopters, like this, that could be a new source of sales growth, but – more importantly – how profitable each of those sales is.
From the enablers to the adopters
AI is probably one of the most important investment themes of my career, but I think we're moving into a different phase.
So far, the market has rewarded the picks-and-shovels suppliers. From here, investors should start asking who's actually making money from all this spending. Who owns the intellectual property? Who has pricing power? And who earns a decent return on the capital they're investing? The sums involved are enormous.
The industry is planning something like 200-220GW of AI capacity, which is almost 20% of America’s current electricity production capacity. The hyperscalers can fund perhaps half of that themselves, and the rest needs outside capital that's getting harder to find. Higher bond yields raise the bar further. Fewer companies can grow earnings fast enough to clear it, which makes stock selection matter much more.
I expect some blockbuster earnings growth and upgrades to forecasts this third-quarter reporting season. Whether investors look past higher rates and the risk of renewed escalation in Iran after the US midterms is another matter.
Concentration is a risk in its own right. Ten stocks are now 40% of the S&P 500. With so much trading driven by passive funds and leveraged retail investment in single stocks, those heavyweights could make for a painful exit when the herd decides to leave.
We're not positioning for some enormous rotation we claim to predict. We're building a portfolio that doesn't need one particular version of the future to be right.
James Thomson is manager of the Rathbone Global Opportunities fund. The views expressed above should not be taken as investment advice.