Second wave: Where does the next AI opportunity lie?

Matt Brennan, Head of Asset Allocation & Market Insight, considers where investors could look for second-round beneficiaries from the artificial intelligence (AI) trend.

The potential for AI to bring about a sea change in ways of living and working has led to soaring share prices and vast capital expenditure plans for an elite group of AI developers, cloud platforms – which host the hardware and store the data – and semiconductor producers. As the market matures, investors may refocus capital in search of fresh opportunities, moving from the suppliers to the appliers: from businesses that supply the technology to those that apply the technology.

Looking for the next opportunity is a recurring pattern when it comes to technological change. The search for beneficiaries of the internet boom, switched from a focus on internet companies and network equipment specialists to retail, logistics and payment systems, for example. Even with the birth of the motor car, second-round beneficiaries arose as the car proliferated, with financial services (insurance), leisure and hospitality (restaurants and motels), road builders and fuel suppliers among the winners. So, where might AI help build revenues, find efficiencies or uncover new knowledge? We will consider equity sectors and other asset classes.

As the market matures, investors may refocus capital in search of fresh opportunities.”

Utilities: Power up?
We’ve all heard about the power-guzzling needs of AI data centres, needs that are likely to keep expanding as more of this infrastructure is constructed and comes online. AI power demand is like to rise faster than generation and grid capacity, requiring significant investment to meet the growing needs of this new technology This buildout focuses attention on electricity grid operators, power generators and energy equipment manufacturers, engineering specialists, utilities, and battery technology businesses. One of the key risks to this could be the potential for stranded surplus capacity. For example, if technological change means that future AI technologies require much less electricity than is currently predicted, assets could be built that become obsolete relatively quickly.

In a related area, the water-cooling that’s necessary for many data centres should boost demand for water treatment plants, while also adding pressure on water supplies.

Industrials: Being productive?
AI could improve production processes, distribution and automation. This could include helping industrial company workflows become more efficient, strengthen quality control, reduce waste and shorten product development cycles. However, the process of upgrading legacy production lines could be disruptive and add significantly to integration and upgrade costs, and as with AI adoption in all industries, there will be risk of software failures and the process of reducing human inputs into workflows means that careful oversight will still be required.

Healthcare: A healthy choice?
The healthcare sector relies heavily on data and specialist knowledge. While AI has already been enabled in medical devices, it is thought it could also help with clinical operations such as triage, information handling and patient scheduling. In the world of drug research, biotechnology and life sciences, it could help to identify targets, and design and analyse trials. AI could also be put to work to in drugs research, helping to speed up the testing process and the overall drug pipeline. At the same time, there are data privacy concerns, system-integration challenges, and the potential for result biases and inaccuracies.

Financial services: Strong finances?
Across the spectrum of financial services there are myriad processes and procedures that AI is could potentially simplify. The technology may help redesign workflows, increase automation, aid with governance, compliance and risk management, for example. These changes could lower administration overheads and improve customer service levels. However, the financial industry is highly regulated, which could make implementation slow and costly, and maintaining customer trust will be crucial.

IT services: Safe and secure?
AI can help businesses detect and block cyber threats, and help secure data and systems, which could boost demand for specialist security platforms and other IT service providers. At the same time, it will also likely enable increasingly sophisticated attacks, adding to the costs and complexities of systems defence. It is also likely that AI could commoditise some aspects of this sector.

Real assets
Real assets encompass everything from infrastructure to real estate and commodities. Investment in property such as data centres, network towers or generation assets could produce long-term revenues. Similarly, private infrastructure funds could benefit from the regulated returns from investing in power grid asset bases, or from owning flexible generation capacity and battery assets that grid operators could tap into when required, to help balance power supply and demand. Key risks to investment in such assets could include technological obsolescence, overrunning construction costs and project leverage, as heavy debt funding tends to be used and the cost of accessing this capital has been rising.

The cyclical drivers for metals, such as copper, and other basic materials could be bolstered by AI-related demand for electrical and transmission equipment and networks. Commodities can offer less reliance on the growth, profitability and financial strength of a single AI company but would add exposure to greater political and geopolitical risks and drivers, as well as a higher level of economic cyclicality.

Credit
Demand for capital for the AI datacentre buildout is leading to vast bond issuance, including investment-grade rated bonds, project debt, private credit and asset-backed finance. Bonds could offer attractive yields, backed by contracted cashflows, and with potentially less volatility than equities. However, investors will need to remain alert to project leverage, refinancing and resale risks, as well as dependence on the mega-cap technology businesses driving this growth. Some of these loans could be viewed as only slightly less-risky technology plays.

In our view, in an increasingly concentrated equity market, looking beyond the AI technology behemoths could improve diversification while keeping some exposure to the technology’s broad growth potential. Many of the diversification options have already attracted investor attention, but significant pockets remain relatively attractive and are not solely reliant on the sky-high expectations that have pushed up AI-related share prices.