Investors have been wobbling on the artificial intelligence trade. Having been convinced that AI would usher in a new golden age of productivity and economic growth, doubts appear to be forming.
Will companies see a return on their vast spending? Will Chinese models pose a competitive threat? Perhaps most importantly, will AI deliver the promised productivity improvements for companies – and at a price worth paying? These are valid concerns, but it is helpful to separate the market response from the reality of AI adoption.
Please remember that the value of investments will fluctuate and returns may be less than the amount originally invested. Tax treatment depends on your individual circumstances and tax rules can change. Chelsea does not offer advice and so if you are unsure of anything please contact an expert adviser.
In the final week of July, there was a dramatic sell-off in semiconductor companies. This was less a reflection of any inherent weakness from this part of the market, which continues to deliver strong earnings, and more a sign of the leverage and hype that had built up.
Separating the winners from losers
Even fund managers that believe in the long-term potential of AI recognised that the single-minded focus on a single area may have been unhealthy. Mikhail Zverev, manager of the WS Amati Global Innovation fund, says: “We saw some signs of unquestioning optimism. We can’t complain, because we’ve benefited from memory semiconductors and photonics. These are key enabling technologies for AI. These are profitable, cash-generative businesses, with valuations justified by their much-improved fundamentals. But we have to be honest with ourselves and recognise that some of this unquestioning enthusiasm, bordering on irrational exuberance, also contributed to the run up in our stocks.”
This early, simplistic judgement on the winners and losers in AI is being challenged, which partly explains the current volatility. Investors are recognising that it is not necessarily a binary question of ‘semiconductors good, software bad’, but that AI benefits will accrue in different ways. For example, Zverev says that Amati is currently digging through the rubble in IT services and software to find companies where the sell-off may have been unjustified.
This discernment is also evident from fund managers looking at larger companies. David J. Eiswert, manager of the T. Rowe Price Global Select Equity fund, says there is a widening gap between companies that can turn higher investment on AI into stronger returns on capital, and those that cannot.
“There is a richer opportunity set for active investors who can distinguish between CapEx that could enhance returns and spending that dilutes them. It is more than a market rotation—it is a shift from concentration to dispersion. In the next phase, returns are likely to depend less on index exposure and more on identifying where capital intensity is creating better economics—not eroding them.”
Alistair Boyle, portfolio manager on Lazard’s Global Quality Growth team, continues to back Microsoft and Accenture as long-term winners from AI. “We expect Microsoft to continue to use its installed base and leadership in the cloud to address secular drivers with a portfolio of high-demand products.” On Accenture, he says: “We continue to believe the company is well positioned as more enterprises move to the cloud, upgrade data, and digitise their business processes as they adopt and integrate AI into their organisations.”
The wider AI ecosystem
For fund managers looking beyond these larger companies, there are plenty of AI-adjacent opportunities that have not attracted the hype of semiconductor and memory names. Zverev, for example, is exploring sensor proliferation. Machines are getting smarter – from robotics, to military drones, to passenger cars. They need to see more and hear more. This is creating a range of companies focused on ‘sensory development’.
T. Rowe Price agrees that the opportunity set is broadening out. “Incremental dollars are increasingly flowing beyond silicon into power, data centre infrastructure, electrical equipment, cooling, connectivity, construction, and services, driving activity across the real economy. In that context, AI is no longer just a technology story; it is becoming a broader industrial and infrastructure investment cycle.”
For the time being, few question that AI will be a major and transformative technology. There have been some set-backs in its rapid growth. Some of the companies that have blazed a trail on AI adoption have pared back their usage on discovering that the costs were far higher than expected. Amazon, for example, saw significant cost over-runs in its early usage of AI*.
Rather than undermining the long-term AI investment case, these challenges highlight how quickly the technology and its economics are evolving. The opportunity is already extending beyond the early semiconductor winners into infrastructure and a wider range of industries, but the benefits are unlikely to accrue evenly.
It makes selectivity increasingly important: identifying companies that can translate AI investment into sustainable productivity gains and stronger returns, while remaining disciplined on valuations and avoiding the sector’s most crowded trades.
