Investors May Be Underrating AI, but Concentration Is the Bigger Risk
Investing·November 13, 2026
Ever since ChatGPT pushed generative AI into the mainstream, it has been hard to read a financial story without the topic appearing. Yet HSBC's global chief investment officer believes many investors are still underestimating how far the technology could reshape the economy and corporate profits in the years ahead.
For portfolio managers, the more urgent question is not whether AI matters but how to position around it. Enthusiasm for the theme has concentrated a large share of stock market gains in a small group of large technology companies, the firms building chips, cloud infrastructure and foundation models. That exposure can look like a clean bet on AI, but it also leaves a portfolio vulnerable if those few stocks stumble, or if the biggest benefits end up flowing to companies and sectors that have not yet been identified.
That is where diversification comes in. Rather than chasing the most obvious winners, investors can look across the wider chain of beneficiaries, from companies that build the hardware and supply the power for data centers to firms in healthcare, finance and other industries that may use AI to cut costs or grow revenue. Spreading holdings across regions and asset classes can also cushion a portfolio if valuations run ahead of earnings.
The argument does not suggest sitting on the sidelines. Ignoring AI altogether risks missing a structural shift that could last for decades, while piling into a single theme can turn a long-term story into a painful short-term loss. The goal is a portfolio that can benefit from the technology without depending on the fortunes of any one company or outcome.
For individual investors, the lesson is familiar but worth repeating. Check how much of your holdings are tied to the same handful of companies or the same theme, and revisit that exposure as the picture changes. Major technologies have a habit of creating winners in unexpected places, and a diversified approach is often the best way to capture that upside without betting everything on the first names that come to mind.
Reporting based on an external source.