Markets will avoid bear territory thanks to the enormous amount being spent on AI, according to Guy Stear, head of developed markets strategy at the Amundi Investment Institute, who told Trustnet that he struggles to see a world in which the global market drops more than 15% before rebounding, provided there are no left-field shock events.
This is because the downside is capped and any setbacks investors experience for the rest of the year are likely to be temporary, he explained.
“Fairly rapidly, people would start saying: ‘Well, the technology is useful, it makes sense and you can do things with it – what are the sectors that will really be able to create products that make money?’ And I think there would be a fairly quick rebound as we start thinking about the next phase of what AI can do,” he said.
That is not to say that investing in AI today comes without risk. Issues that could cause markets to slump include the IPOs from Claude owner Anthropic and ChatGPT’s OpenAI struggling to get the valuations they crave or the frothy prices of AI leaders already on the market start to be challenged.
In particular, he has “sympathy with the argument” that companies solely linked to large language models are vulnerable, noting that the valuations of some have become stretched.
“If these things perform poorly, then what you [will see is] a revision of people's thoughts about the profitability of AI. That's certainly bad for the sector but it also means a reset of the entire equity market,” said Stear.
Another concern is the fear that hyperscalers are spending too much and capital expenditure has got out of hand. This is a possibility if AI ends up purely being about large language models, but Stear’s base case is far more bullish than this.
“It isn't simply large language models, because for AI [to work] in the physical world, [for example to run things like] drones, we're going to need even more connectivity and even more data-crunching ability. We will rapidly realise that we need even more data centre rollouts, so capex is going to continue,” he said.
Amundi’s growth forecasts are predicated on AI taking off. The firm’s capital markets assumptions look at what capex levels are reasonable for the next three to five years.
“They’re big numbers,” said Stear. Indeed, the firm’s report suggests a 10% growth in capex versus the US economy, taking the amount spent by hyperscalers on AI to around 17% of US GDP.
“We haven't seen that for decades. It's really a step change. And we strongly believe that is going to be true,” he said.
Whether markets fall or not, one area that he said could continue to disappoint is software. The sector has taken a hammering this year as investors have been quick to price in AI taking over from current technologies.
Stear said the sector has “a lot more vulnerability”, adding that he can understand why some investors believe earnings multiples are still too high.
“The one thing we've learned from AI is that it is very good at writing software. So the intellectual property of software companies is not as valuable as it was because it's easier to create,” he said.
Eleanor Ingilby, head of high net worth at atomos, said there is merit to the idea that markets are well supported thanks to the growing number of retail investors, who are more notoriously willing to act decisively.
“As people engage more, we're seeing quite sharp rebuys as soon as there's a drop in the markets,” she said, which isn’t the case with typically slow-moving institutions.
However, she noted that it is hard to make predictions and investors should heed the lessons learned from events such as the Covid pandemic and financial crisis.
“What if the cause of the market crash was AI? What if, for example, some AI model runs wild? What if Anthropic gets hacked and suddenly it takes down major energy infrastructure?”
In these cases, the market could fall much further than expected – although these events would be far beyond Stear’s standard market correction hypothesis.