Investing in AI: Similar or different to other technologies?

Artificial intelligence (AI), which has been around for decades in various forms, has burst into the mainstream in recent years, including applications like ChatGPT and autonomous driving software. Given how the technology can automate tasks and substitute machines for humans, investors are now thinking about the role of AI in increasing output, reducing costs and changing the nature of business — and how this might affect industries and companies. In this article, we'll explore some of the things to consider about investing in AI, including how AI might be similar or different to investing in other technologies.

Company roles in the AI ecosystem

To get started, let’s consider the roles that companies play in the AI ecosystem:

Developers: Companies that develop the software and standards of AI across a wide range of markets, such as technology firms.
Builders: Companies that build the hardware or physical infrastructure for AI, such as chipmakers and data centres.
Enablers: Companies that enable others to use AI, such as consulting and professional services firms.
Integrators: Companies that play all of the above roles.
Deployers: Companies that don’t play any of the above roles but that could be major users of AI in their business, such as banks or retailers.

These roles can mean different things to investors. If AI is similar to other technologies, it's possible that a handful of companies will become big winners as developers of the software or builders of the hardware, like software developers and chipmakers in the personal computer industry. It's also possible that only a few companies will emerge as integrators, like the current market for internet search technology. In these AI categories, investors have already seen breathless predictions and soaring valuations, given the potential for the winners to dominate global markets. Yet this heady excitement may be less applicable to deployers — companies that adopt or adapt the technology. To illustrate this point, think of the dot-com boom of the early 2000s. Only a few companies turned out to be leading-edge developers of internet technology, while the vast majority of all other companies simply used the technology in their businesses.

Opportunities and risks in AI investing

Opportunities and risks, always present in investing, loom larger in early-stage technology like AI. Here are some to consider.

Picking the winners

When it comes to the creators of a new technology — the developers, builders and integrators — it can be notoriously difficult to pick the winners. A key dilemma is whether to invest in start-ups or established companies. Start-ups may offer pure-play exposure to AI, although with much higher risk than investing in a large company. Established companies may be able to use their market dominance and wealth of customer data to capitalize on AI, although with the risk of losing existing lines of business. AI, as a theme rather than a product line or business segment, may be difficult to measure through revenue or earnings on a company's income statement. Investors may have to sift through analyst reports and earnings calls to get a sense of how AI is impacting the business. Perhaps AI can help with that.

Focusing on results

When it comes to the users of a new technology — the deployers — AI could be similar to other capital investments. Success is ultimately about implementation and results. For AI to add value for investors, a company would have to deploy AI better than its competitors and with results that are better than expected by analysts. In this sense, AI may be no different than other technologies like desktop or cloud computing. The technology may increase capital expenditure and reduce operating costs over the long term — but competitive pressure may result in little or no improvement in market position for many companies. Investors should expect a competitive response when companies deploy AI.

Deciding how much to pay

Based on valuation theory, the value of a company's stock is the present value of its expected future cash flows. This involves projecting cash flows far into the future and discounting them to the present with an interest rate that accounts for the risk of actually receiving those cash flows. A common tendency in growth investing — the "growth trap" — is for investors to overestimate the cash flows and underestimate the discount rate or risk. This tendency is likely to be more pronounced in the hypergrowth world of AI. Even wonderful companies can make woeful investments if investors pay too much for future cash flows. There is a risk of AI champions being "priced for perfection", where anything less than outstanding performance against high expectations could result in significant declines in market value.

Are you an AI believer?

Thinking about opportunities and risks can help to inform an investor's degree of conviction about the prospects of AI and their ability to pick winning investments in this market. Here are several ways to invest in AI, ranging from a higher to a lower level of conviction.

Individual stocks: This is the highest form of conviction to invest in AI, given the difficulty of identifying which companies will develop, build or integrate the technology in such a way to dominate global markets, or which companies will deploy the technology better than their competitors.
Thematic funds: Investors who have some conviction in AI but not enough to pick individual stocks may consider an AI-themed fund. However, fund managers may have wide discretion about how they define the theme and select the companies.
Broad market funds: This is the lowest form of conviction to invest in AI. If AI turns out to be an important trend across the economy, AI champions would be well represented in a broad market fund, including a global market fund for investors who don't want to predict which countries or regions will be the best AI performers.

Key takeaways

Companies can play a wide range of roles in AI. To think clearly about investments in AI, investors can consider whether a company is a developer, builder, enabler, integrator or deployer of AI technology.
AI technology is at a very early stage. This makes for considerable opportunities and risks, including picking winners among developers, builders or integrators; focusing on results for deployers; and, overall, deciding how much to pay for potential earnings that could extend far into the future.
Investors may have varying degrees of conviction about the importance of AI and their ability to select winning investments in this market. Ways to invest in AI, ranging from a higher to lower level of conviction, include individual stocks, thematic funds and broad market funds.