Harry Topping
In recent years, Artificial Intelligence (AI) has become a key driver of markets and one of the hottest topics across the globe. Since the launch of ChatGPT in 2022, the world’s biggest companies have scrambled to acquire the microchips and computing power to put themselves at the cutting edge of this transformative technology. But what exactly do we mean when we talk about AI? How does it work and what are the applications? And how can we try to gain exposure to this theme within the Equilibrium funds?
In its simplest form, AI works by recognising patterns in old data to help us answer questions and solve problems in the present. This can be in the form of recognising patterns to help us forecast data, or even how ideas and words fit together in order to give us human like text and voice responses to queries. Today we have access to ‘Narrow AI’ – systems designed to perform specific tasks. ‘General AI’, capable of performing any task that would usually require human intelligence, remains only theoretical.
If we follow through the idea of recognising patterns, we can identify some areas where AI is most prevalent today. The first, and most often used, is the ever frustrating AI chatbots used by companies to deal with complaints. If the AI can recognise patterns in how common issues have been solved in the past, it can point complainants toward the answers they require. Companies now claim 85% of complaints can be handled with no human intervention.
Another less recognised area is journalism – 50% of newly published articles in 2025 were written by AI. This might seem like a sector insulated from technology, but most articles are relatively predictable collections of facts and prose, lending themselves to imitation through pattern recognition.
From an economic perspective, the great hope for AI is that it can reignite productivity growth in the developed world, or growth in output per worker. Chart one shows productivity in the UK between 1990 and 2025. From 1990 to 2008, productivity grew at 1.8% per annum, which compares to just 0.86% from 2009-19 and -0.28% from 2022-25. This is certainly a source of economic woes as nations have relied on increasing the number of workers for growth rather than improving their output. On the bright side, there is real hope that AI can help automate tasks and allow workers to provide more value to the economy in the coming decades.
Potentially transformative technologies always present both investment risks and opportunities. One way to gain investment exposure is through the ‘Hyperscalers’ – large US tech companies supplying the computing power needed to run AI models via the cloud. Microsoft, Google and Amazon are prime examples, who also have access to proprietary data for training bespoke models. Then there are semiconductor companies producing the chips needed to train and run AI models. Nvidia is currently the world’s largest company with a near monopoly in designing best in class chips, while Taiwan Semiconductor Manufacturing Company (TSMC) dominates high-end manufacturing.
While earnings have grown strongly, valuations have become expensive – some might claim these companies are now ‘priced for perfection’. Recent deepening ties between AI related companies also give cause for concern. Chart two below gives a flavour of how interlinked some of these companies have become. OpenAI, the parent company of ChatGPT, has pledged approximately $1.5 trillion in future expenditures, despite currently generating only $14 billion in annual revenue.
History would recommend approaching these lofty valuations and interconnectedness with caution. There are, however, other ways to access this theme at more reasonable prices. One is through companies supplying data centres with the huge amount of power they need the largest facilities under construction will have consumption comparable to San Francisco. Another is providers of cooling systems that allow data centres to maintain constant temperatures regardless of demand. Finally, if we see the promised productivity gains, it will likely be new companies with AI at their core that displace incumbents and become the winners of tomorrow. Consider that the World Wide Web emerged in 1989, but Amazon wasn’t founded until 1994 and Facebook only in 2004.
We are in the foothills of this new technology, which will undoubtedly change the economy in ways we can’t yet understand, but we must remember that transformative technology doesn’t translate into investment returns for all. We will continue to be evidence based investors, applying a healthy level of scepticism in trying to identify where future investment returns might accrue.
This blog is intended as an information piece and does not constitute investment advice.
If you have any further questions, please don’t hesitate to get in touch with us on 0161 383 3335 or by reaching out to your usual Equilibrium contact.
