The Viewpoint: Testing the AI trade

Coming back from four weeks of paternity leave in June I was looking forward to the relative tranquillity of uninterrupted equity research and portfolio construction during the summer months. The market has had other plans, with capital flowing into and out of various geographies, sectors or styles at an incredible pace, particularly during late July. 

In plain sight

Remarkably, the news flow that has triggered these moves hasn't been particularly surprising – we know China is working frantically to compete in semiconductor equipment, we can see the threat to the economics of frontier model providers from open-source (commoditisation is taking place in front of our eyes) and it is widely reported that the AI ecosystem is becoming increasingly interdependent. And against that backdrop, the AI debate is becoming a contest between forces that rely on each other but are far from aligned.

The news flow that has triggered these moves hasn't been particularly surprising

For those who like a fantasy drama, our Game of Thrones analogy is starting to write itself. House Targaryen are the fire-breathing memory dragons that power the AI revolution, where the greatest threat comes from within. Will the hyperscaling House Lannister ‘always pay their debts’? The legacy semiconductor companies of The North, proud and historically dominant, are aggressively trying to adapt to the change in seasons. The Night’s Watch of TSMC and ASML continue to hold the civilised world together but if the Wall falls, everything ends. And finally, the White Walking AI Bubble, a looming, existential threat that half the realm believes is a myth and the other half is terrified of.

Is Winter coming?

The crux of the AI debate is whether the technology will deliver a real return on investment for the firms spending the big bucks on it – and those spending others' big bucks on it for the promise of future riches. A widely cited back of the envelope calculation compares the roughly $1 trillion poured into AI infrastructure against the estimated $150 billion in annual recurring revenue generated so far. While a 15% top-line yield on this investment might look reasonable on paper, the maths quickly breaks down under hardware depreciation and typical operating margins, which leave total industry revenue well short. Regardless of the future efficiencies which we will see at the hardware and model layer, one thing is clear: for this buildout to deliver acceptable returns, industry revenue needs to climb by orders of magnitude.

One thing is clear: for this buildout to deliver acceptable returns, industry revenue needs to climb by orders of magnitude

Given the wide range of outcomes, let’s move away from the numbers and focus on the narrative. Up until now there has been a cushion – the cash flow of the most successful businesses that we have ever seen has supported much of this investment (the degree to which this continues to be the case is debatable…). Regardless of how this investment is funded, as more and more capital is committed, the stakes get higher and higher, and the path becomes narrower and narrower. That’s when focus shifts entirely to demand and the need for ‘tech diffusion’ to drive revenues. Put simply, you have to win every consumer or enterprise tomorrow. ‘You need your grandma to start vibe coding’.

What does an investor do in this market?

We do not intend to take a view on whether AI is a bubble or the greatest investment opportunity in decades (in reality these are not mutually exclusive) but while it paid to look far and wide for these opportunities in 2025, we believe that more recently, the clearer investment cases are right in front of us. This year our team has continued to dig deep into the supply side for investment opportunities – exploring constraints in fabrication, grid infrastructure, metrology and even critical insulating resin films (this company happens to also be world-famous for making Umami seasoning...). But for now it is the more recognised – possibly even boring – leaders across the ecosystem where the price you pay for potentially materially higher future profits appears most attractive to us. Fortunately, many of these more established companies recognise where they sit in their corporate life cycle and pay reasonable dividends. It is some of these stocks that we topped up on as the market sold off in late July. 

The sell-off wasn’t just an opportunity to slightly reduce our technology underweight, it also offered a useful test of how AI-related news flow is now feeding into the broader market. Until recently, it has been harder to gauge how and when this news flow would affect market behaviour. Last month offered a glimpse into what an increasingly retail-driven market structure and the positioning of its ever more levered participants can mean for share prices when the narrative changes.

The news flow that has triggered these moves hasn't been particularly surprising

Our global equity income strategy has come through this recent period in reasonable shape, although we would not draw too much from a short period of market volatility. Having outperformed over the last 18 months, in a difficult environment for equity income funds, we have continued to question whether this has been driven by stock selection, as our attribution data suggests, or whether we were carrying unintended exposures that our risk tools or judgement had yet to uncover. In that context, the late-July market stress was a useful test. Our global equity income strategy performed well as the market rotated towards companies with more perceived visibility and less stretched valuations, before giving up a material portion of that relative performance as the trend sharply reversed. Overall, the pattern of returns has been broadly consistent with how we would expect the strategy to behave in this environment.

The longer view

In the meantime, our investment process continues to hum. During the month we initiated positions in an unloved US infrastructure software firm and a loathed Danish continence company – both putting it mildly. These are not fashionable areas of the market, but that is part of the appeal. In both cases, we see businesses with durable franchises, scope for better execution, and valuations that already discount a fair amount of disappointment. The former also includes a leading quantum computing division that could prove highly relevant as the limits of today’s frontier AI models become clearer. Alongside regular portfolio fine-tuning, we believe that these additions leave us with a broad set of return drivers and what we believe is the right balance between participating in strong equity markets and defending capital when conditions become more difficult.

Now back to watching UAI (un-artificial intelligence). In the last few weeks our daughter has figured out how to grip toys, express joy and boredom, and improve fuelling efficiency – all problems that humanoid developers have been wrestling with for years, if not decades. Perhaps AI in the physical world isn’t quite what it’s made out to be. Or is she just an early developer? 

For professional investors only. This material is not suitable for a retail audience. Capital at risk. This is a financial promotion and is not investment advice. Past performance is not a guide to future performance. Reference to any security is for information purposes only and should not be considered a recommendation to buy or sell.

The value of investments and any income from them may go down as well as up and is not guaranteed. Investors may not get back the amount invested. Portfolio characteristics and holdings are subject to change without notice. The views expressed are those of the author at the date of publication unless otherwise indicated, which are subject to change. Forward looking statements are subject to certain risks and uncertainties. Actual outcomes may be materially different from those expressed or implied.

Investment Risk: The value of investments and any income from them may go down as well as up and is not guaranteed. Investors may not get back the amount invested.

EPM Techniques: The Strategy may engage in EPM techniques including holdings of derivative instruments. Whilst intended to reduce risk, the use of these instruments may expose the Strategy to increased price volatility.

Exchange Rate Risk: Investing in assets denominated in a currency other than the base currency of the Strategy means the value of the investment can be affected by changes in exchange rates

Liquidity Risk: In difficult market conditions the value of certain strategies may be difficult to value and harder to sell, or sell at a fair price, resulting in unpredictable falls in the value of your holding.

Emerging Markets Risk: Investing in Emerging Markets may provide the potential for greater rewards but carries greater risk due to the possibility of high volatility, low liquidity, currency fluctuations, the adverse effect of social, political and economic instability, weak supervisory structures and accounting standards.

Counterparty Risk: The insolvency of any institutions providing services such as safekeeping of assets or acting as counterparty to derivatives or other instruments, may expose the Strategy to financial loss.

Charges from Capital Risk: Charges are taken from the capital of the Strategy. Whilst this increases the yield, it also has the effect of reducing the potential for capital growth.