AI is a bet we’re forced to make

Sometimes I wonder if the world outside Silicon Valley really wants AI. Yes, the technological advances in this area are staggering. Watch AI tools and agents analyse information, write text, produce photorealistic images and video, and – perhaps most disconcertingly of all – directly operate other computer programmes via text and visual interfaces designed for humans. It is impossible not to be impressed. At the same time, many people say that AI is being forced on them (they notice the proliferation of AI functions that can’t be turned off in all the software we use), they feel no need or desire to try AI, and they think fears about the risks and social consequences are being ignored.

None of this means that AI will not be useful or will not take off – which are not quite the same thing. There were many cynics about the economic value of the internet in the early days – there are few these days. Social media took off like a virus – and now there are growing concerns about the harms. However, the sense that big tech is pushing us harder to use AI feels true. This reflects the sheer amount of money that these companies are funnelling into it, and how much they need it to pay off.

Heavy investment in AI

Julian Bishop and James Ashworth of Brunner Investment Trust (LSE: BUT) summed up the reasons to be cautious well at an event last week. Look at the amount of money being invested – maybe $400-$450 billion this year. Bear in mind that companies say they plan to keep doing this until 2030. You’re looking at $3 trillion in capital expenditure (capex). Look at the revenues being reported by OpenAI and Anthropic as a partial proxy for core AI-based revenues – maybe $20 billion this year for the two. This is growing fast, but it would need to grow a long way further to earn a decent return on capex of this scale. The capex intensity of this stands in sharp contrast to how capex-light the tech giants used to be, which has implications for free cash flow. Note too the way that this capex flows through the market, from energy infrastructure to real-estate developers building data centres. “Large chunks of the market’s ongoing strength are contingent on AI working out,” says Bishop. (See also the chart above from Joel Suss in the Financial Times, implying that AI spending is now dominating US investment.)

Start your trial of MoneyWeek today

Get unparalleled financial insight, analysis and expert opinion you can profit from.

Start your trial

Sign up for MoneyWeek’s free twice-daily newsletter.

Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.

Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.

Sign up

A capex bubble can have long-term benefits, even if investors lose out – think of canals and railways. Yet here the money is being spent on short-life assets, notes Ashworth. Nvidia’s costly chips may be in use for five years or less.

On the plus side, the US tech giants are highly cash generative. It may not be irrational in the long term for them to spend heavily now to defend their positions against potential threats. The greatest froth is in private markets, say Bishop and Ashworth; they hold some of the mega caps. Still, they are clearly on the alert for signs that the cycle is turning, and this feels right. We cannot ignore the idea of an AI revolution, but with the tech giants making up 35% of the S&P 500, there is a lot of downside if it evaporates.

This article was first published in MoneyWeek's magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.