Long Read  

AI will take time to show a return on investment

“But we question how it is going to work and whether it will generate sufficient return on capital.” This is a major question for all involved in the AI supply chain. 

Cutler prefers to tap into the trend through utilities: “It’s a wonderful shot in the arm for the critical energy infrastructure theme.”

Article continues after advert

He says only two out of the five combatants are likely to come out on top: “Do you bet on all five, or do you bet on the ammo providers? That’s where we are.” He is focusing on D-ram providers such as Samsung, and chip manufacturers such as TSMC. He is less positive on Nvidia, which designs chips, but relies on TSMC to manufacture them. 

Excessive spending

James Harries, manager of the Trojan Global Income fund, is also a sceptic: “We’re not going to say that the Magnificent Seven are poor companies. Absolutely not. But I do wonder whether the CapEx boom related to AI is a classic arms race. Companies don’t want to be left behind and are spending on an almost non-discretionary basis.

“They are spending like there is no tomorrow to build out the infrastructure in the hope that there will be utilisation for AI. This could lead to over-capacity. That could be quite material not just for AI and semiconductors, but for technology companies that are spending a lot of money, the returns of which are still indeterminate.” 

Even AI specialists say the next stage will be more complicated.

Chris Ford, co-manager of the Sanlam Global Artificial Intelligence fund, says that while the early priority had to be the building of infrastructure, “companies can’t sell products and services until that infrastructure has been built”. 

Those products and services will take time to emerge. When they do, they could be extremely disruptive and high growth, but the path will not be linear. 

Markets may already be reflecting this reality. The relative growth of the large AI names versus the rest of the market has slowed, with even the unloved FTSE 100 outpacing the Nasdaq over the past three months.

The Nasdaq is up 4.43 per cent over three months, while the FTSE 100 is up 9.78 per cent. This is very short-term, but it shows that the market may be starting to change its view. 

Vast investment has gone into AI infrastructure, but it is not yet clear how AI will be deployed.

Where are the life-changing use cases for it? These will almost certainly emerge, as they did with the internet, but it will take time, and companies may not see a return on their upfront investment in the short-term.