The Problem With Chasing AI

It is hard to open a financial news site right now without seeing another headline about artificial intelligence. The enthusiasm is understandable. AI is reshaping how businesses operate, how decisions get made, and how capital gets allocated. Over the next decade, it will likely create some of the most valuable companies the world has ever seen.

So why aren't we simply buying everything in its path?

Because price matters. It always has, and AI doesn't change that.

When a theme captures the imagination of the market, capital floods in fast. Valuations stretch. Companies that are adjacent to the trend — sometimes only loosely — get pulled along for the ride. Investors tell themselves the story is so big that normal rules don't apply. That's precisely when discipline becomes most important, and most difficult.

We believe AI is a genuine, multi-decade productivity shift. We own businesses that are direct beneficiaries: semiconductor companies, hyperscalers, and platform businesses that are embedding AI into products that billions of people already use. But our AI exposure doesn't stop there. Many of the businesses we own across other industries are quietly using AI to streamline their operations, reduce costs, grow revenue, and expand margins. They are not AI companies in any headline sense, but they are becoming meaningfully better businesses because of it. This is, in our view, one of the more underappreciated dimensions of the AI story, and one that doesn't require paying a premium to access.

These are real businesses generating real cash, and we think the long-term earnings power justifies ownership at sensible prices.

But sensible prices is the key phrase. We are not in the business of buying a good story at any price and hoping someone pays more later. That's speculation, not investing. The distinction sounds simple, but it gets blurry when markets are moving fast and everyone around you seems to be making money.

The other thing worth saying clearly: AI will also produce a lot of losers. Not every chip company, data centre operator, or software platform will emerge as a winner from this cycle. There will be overbuilding. There will be hype cycles that correct sharply. Some businesses that look indispensable today will be commoditized faster than anyone expects. Picking through that carefully, rather than owning the theme wholesale, is where we think the real work happens.

Our portfolios are not AI portfolios. They are portfolios of high-quality businesses, some of which happen to benefit meaningfully from AI. That distinction matters more than it might seem. It means we are not one bad earnings report away from a thesis falling apart. It means we can hold through volatility without second-guessing the core logic. And it means our clients are exposed to the upside of a powerful technological shift without betting the entire portfolio on it playing out exactly as the market currently expects.

AI is not a reason to abandon discipline. If anything, it's a stress test of it.