Инвесторам необходимо оценить скрытую концентрацию портфелей в искусственном интеллекте

Figuring out how much artificial intelligence exposure a portfolio contains is becoming increasingly complicated as the investment theme expands well beyond the obvious names. “If you own a stock like Nvidia, then obviously you could draw a line and say, ‘I have some AI exposure,’” says Spencer Morgan, director of portfolio strategy at Purpose Investments Inc. in Toronto. But determining how much AI-adjacent exposure index funds have is harder. In a recent report, Purpose Investments attempted to put a number on that exposure, developing a framework that estimates roughly 45 per cent of the S&P 500 is invested in AI-related companies. Add in companies in non-technology sectors affected by the theme and that rises to nearly half the index. “If you own the market, you have made a large, concentrated AI allocation just by being allocated to the largest equity market in the world,” the report states. The goal isn’t necessarily to own more or less AI, but to understand where that exposure is coming from and whether what looks like a diversified portfolio is actually concentrated in the same investment theme, says Michael Greenberg, head of Americas portfolio management at Franklin Templeton Investment Solutions in Toronto. “We’re not saying this is like the [2000] tech bubble and you need to run for the hills,” he says. “We actually believe this is a theme that’s got some legs and it could last quite some time. It’s just all about how many eggs do you want to have in that one basket.” Part of the challenge is that AI exposure doesn’t fit neatly into the traditional classifications investors may use to understand their portfolios. “There is no GICS [Global Industry Classification Standard] sector that is AI. There’s no specific definition currently. And that’s really the crux of the problem,” Mr. Morgan says. Instead, AI exposure can stretch from the most obvious beneficiaries – such as semiconductor companies, hyperscalers and model developers – to robotics and automation, software and the infrastructure needed to support the massive AI buildout, including utilities and power producers. That means knowing how much AI is in a portfolio is only the first step. Investors also need to understand what kind of exposure they own, as different types carry very different risks. For example, Purpose estimates emerging markets have an AI exposure score of about 24 per cent, compared with about 10 per cent for Canada. But the exposures have little in common. Emerging markets are heavily weighted toward the “picks and shovels” AI trade, with foundries, memory and packaging companies, while Canada’s exposure is largely application software. Mr. Greenberg says investors need to look at their portfolios through several different lenses to uncover those risks. “If you’re just looking at any one of these layers in isolation, you’re probably not getting the full picture,” he says. He suggests starting with geography and then looking at sector and industry exposure before going deeper into factors such as value, growth, momentum and market capitalization. An investor could, for instance, hold U.S., South Korean and Taiwanese investments and appear to have geographic diversification. But all three markets can have significant exposure to the same parts of the AI trade. The same issue can arise when investors own several ETFs. Someone holding an S&P 500 ETF, a Nasdaq ETF and an AI thematic ETF may appear diversified based on the number of funds they own, but the underlying holdings can overlap considerably. “You may think you’ve got some diversification there, but you probably don’t have a whole lot,” Mr. Greenberg says. The companies that have benefited most so far have largely been AI “enablers,” he says, including semiconductor makers, hyperscalers and electrical-generation companies involved in building AI infrastructure. The next phase could be much broader as companies become AI “adopters.” That could include retailers, restaurants or industrial companies using the technology to improve productivity, manage inventory or reduce waste. For investors heavily exposed to the first group, Mr. Greenberg says diversification could mean looking to other geographies, sectors and investment styles rather than simply adding another AI-related investment. Valuations are also a consideration, especially after the strong run in some AI-related stocks, which makes it even more important for investors to understand where their exposure lies.