Hedging against AI exuberance

In August 2018, Apple became the world’s first company with a trillion-dollar valuation. At the time, it seemed hard to believe any company could ever be worth such an unimaginable sum. Less than eight years on, there are 15 companies in the trillion-dollar club with a combined valuation of $36 trillion, $4 trillion more than the entire GDP of the US last year. 

In just the past three months, AI hardware companies Samsung, SK Hynix, and Micron joined the illustrious ranks, and three more are expected this summer via US flotations, again all AI related. These are Anthropic, the developer of Claude, the large language model (LLM); OpenAI, the developer of the ChatGPT LLM; and SpaceX, ostensibly a space technology business, but one that also happens to be developing an LLM in the form of Grok. 

SpaceX is seeking a valuation at flotation of $1.8 trillion, which would make it the highest valued IPO of all time, despite it making a net loss last year of $5bn on $18.7bn revenue. It’s successful satellite business, Starlink, made over $4.4bn in operating profit but the huge sums the company is pumping into AI development dragged the performance underwater. OpenAI has also yet to turn a profit, while Anthropic did pull in over half a billion in the past quarter, although it also warned that its huge infrastructure investments mean it won’t become cash positive until 2028.

If these valuations seem excessive it’s because some of them probably are. Unlike the many internet companies on wild valuations at the start of the century, these businesses are generating substantial revenues and are all undergoing explosive growth. Massive investment into AI is real and the technology does have the potential to fundamentally transform the way we live and work, leading to huge rewards for the winners in the AI race.

The issue with valuations such as these is that they assume all these companies will be the winners. However, competition is fiercest where rewards are greatest, and there are countless other companies out there seeking a piece of the AI pie. As with any industry there will ultimately only be a handful of market leaders. These valuations also appear to make many other assumptions, such as that the current boom in AI spending is sustainable, which is far from certain; or that LLMs can generate strong profit margins, which is unproven. The only thing that is certain is that there will be many bumps in the road on the AI journey. 

It would be nice to think these issues exist only in the abstract, but you don’t have to be an active investor in AI to be involved. Most default workplace pension funds in the UK invest heavily in global equities, often following benchmark indexes. These new mega-cap IPOs are being fast-tracked onto these indexes within days which means you may very soon be heavily exposed to these stocks through your pension.

You may view this as a good thing, depending on your expectations for AI, but if you’re a relatively risk-averse investor it might be wise to take this as an opportunity to review your pension allocation. The UK market has a lot of defensive companies at sensible valuations that could act as an excellent hedge. As the US market becomes increasingly frothy, financial communications professionals must make the case for the benefits of investing in UK plc.

In short, the current exuberance for AI may still have further to run, but three near simultaneous trillion dollar plus IPOs certainly feels like a top-of-the-market event. When the music stops the correction could be painful and it would seem sensible not to be too exposed.