HAMISH MCRAE: There is a healthy dose of skepticism about AI

This is the so-called Great Rotation, which began in earnest last week. What it describes is an investor shift away from US high-tech giants, especially those seen as more likely to benefit from artificial intelligence, and toward companies in core sectors with less ambitious price-earnings ratios.

On Wednesday, the Nasdaq index of high-tech companies fell 3.6 percent, its worst day since October 2022.

While the group’s two most valuable companies, Apple and Microsoft, are still worth more than $3tn (£2.3tn), Nvidia, seen as the biggest beneficiary of the generative AI revolution, has fallen below that lofty benchmark. The market recovered somewhat on Friday, but it’s been an unpleasant week.

Two things seem to be happening: one is that a shift in investor sentiment seems to have begun that everyone knew would happen sooner or later (from the hope of growth to the reality of the need to make profits).

The other is that it is increasingly clear that applying AI successfully in practice will be expensive, complex and unpredictable, so it is not entirely clear who will make money from it.

The first issue is familiar. There is always a tension between growth and value. If you choose growth, you accept a low annual return in the hope that the increase in capital value will more than compensate for it. If you prefer value, you pay fairly high dividends in the acceptance that you will never make a fortune from rising share values.

It’s a choice people make for all sorts of reasons: their age, temperament, need for dividends, and so on. But the hard truth is that over the past decade growth has won out. Portfolios of fast-growing companies have achieved higher total returns than those of the highest dividend payers, and because the US market is dominated by high-tech giants, it has grown faster than those in the UK and Europe.

So far, those of us hoping for a change in attitudes toward value have been wrong. But we also know from history that all bubbles burst, so the question worrying investors on both sides of the Atlantic is whether the jitters of the past few days are a harbinger of a broader reappraisal of investment strategy.

Perhaps this is the beginning of a period in which value investing becomes fashionable again.

As for the AI ​​revolution, it is quite clear that it will bring about huge improvements in the performance of service industries. This is already starting, but the problems are also becoming apparent.

One of them is the energy demands of the enormous computing power that AI needs to function. There are cases where coal-fired power plants had to be restarted to keep AI servers running.

Another is the capital investment required to train and develop the models to make them reliable and useful.

Then there is the question of how to apply it practically. How do you save money by using AI instead of people? And what do you do when you find that service quality has declined because you got rid of the wrong staff? There are many worrying privacy issues.

The same is always true in the early stages of a technological revolution. We know that the apparent beneficiaries may not get the expected results and that the real winners may not yet exist.

We saw all that with the dotcom bubble that burst earlier this century, and while the hyperbole is less extreme now, that episode remains a warning to us all.

I actually think what has happened in the last few days is extremely healthy. It is always sensible for investors to ask themselves whether a strategy that worked very well (in this case, backing big US tech) may have come to an end.

And it is healthy, in a more specific sense, that we are sensitive to the drawbacks of AI as well as its potential.

I think it is the Holy Grail that will transform the productivity of service industries, but as in all revolutions, there will be losers and winners, and let’s be frank about that too.