Welcome to ETF IQ, a weekly newsletter dedicated to the $14 trillion global ETF industry. I'm Bloomberg News reporter Vildana Hajric, filling in for Katie Greifeld. In the already buzzy world of ETFs, nothing could have been more electrifying this week than the launch of an AI-powered ETF that’s promising to harness the brainpower of the investment world’s most illustrious minds. A chatbot-powered fund from fintech startup Intelligent Alpha aims to spit out investment recommendations inspired by the thinkings and doings of Warren Buffett, Stanley Druckenmiller, David Tepper, and others. The fund — the Intelligent Livermore ETF, which is fed data and writings from the Wall Street dignitaries — might end up buying companies that the investors themselves may not necessarily be holding currently. Warren Buffett, chairman and chief executive officer of Berkshire Hathaway Inc. Photographer: Christopher Goodney/Bloomberg But while Intelligent Alpha’s founder Doug Clinton said he has had success beating the S&P 500 with this strategy and others, AI ETFs more broadly have posted lackluster performance of late — even as the theme remains the hottest of the year (no offense to Jay Powell, the Olympics and Charli XCX). Of the 16 AI-centered ETFs in the US tracked by Bloomberg Intelligence, just one is outperforming the S&P 500 this year: the Franklin Intelligent Machines ETF (ticker IQM). The fund has returned 22.5% while the stock index has gained 20% as of Thursday’s close. Further, only one — the Global X Artificial Intelligence & Technology ETF (ticker AIQ) — has seen meaningful inflows, taking in more than $1 billion this year. That’s followed by a $117 million haul for the Roundhill Generative AI & Technology ETF (CHAT). The rest have seen tiny inflows or outright outflows year to date. Clinton said that many of the other AI-centered ETFs tend to rely on traditional machine-learning techniques and might not yet be incorporating LLMs like his fund. “That limits those strategies to the crowded market of quantitative insights,” he said. AI funds can be broken up into two broad categories: those that hold AI companies, and those whose investments are driven by AI. Some of them can be thought of as “black-box models that in many cases have insane turnover and effectively wind up chasing their tails,” said Ben Johnson, head of client solutions at Morningstar. In other words, constantly switching up which companies an ETF holds based on an AI’s recommendation could be detrimental to performance. When it comes to flows, it might also be the case that owning an S&P 500-tracking fund gives investors enough AI exposure anyway, and so they find no need buying AI ETFs. That’s because the big AI behemoths, like Nvidia Corp., are already part of S&P 500 funds. “While AI is more intelligent now than ever, there doesn't seem to be a big selling point behind an AI-driven ETF,” said Roxanna Islam, head of sector and industry research at TMX VettaFi. “Investors can get rules-based index investing for relatively cheap, and on the active-ETF side there are plenty of portfolios managed by actual star investors rather than a replica of one. With a proven track record, these AI-driven ETFs could catch on with investors, but we have not seen that occur yet.” |