Traders work at the New York Stock Exchange on May 28, 2026.
NYSE
Assets in leveraged ETF nearly doubled over two months, as investors scrambled for maximum exposure to the artificial intelligence trend.
“Investors increasingly use leveraged instruments linked to the ‘AI-trade’ – assets in leveraged ETFs linked to AI and Tech themes have surged, with exposures in the US and Korea/Taiwan rising sharply- more than doubling in recent months,” wrote Goldman’s Christian Mueller-Glissmann in a note.
Goldman cited data from EPFR tracking 573 leveraged U.S. equity ETFs, 52 ETFs on South Korea and 11 on Taiwan. The total net assets for leveraged equity ETFs on U.S. equities doubled in just two months, climbing to $84 billion end of May this year from $39 billion in April. Leveraged ETFs on South Korea and Taiwan also surged to $43.1 billion from $17 billion during the same period.
The rise reflects the growing appetite of AI-driven stocks, but also raises a red flag that the rally may be partly fueled by money flowing into these risky funds. Leveraged ETFs use derivatives to provide daily returns double and even triple the return of certain indexes and stocks.
There’s risk that the money flowing in could reverse as, or even more aggressively if there’s a pullback in the AI trade as investors face big losses in the funds.
“It’s not surprising investors are reaching for leverage given the extent of the mania happening in AI- a common behavior during all bull markets”, says Adam Crisafulli, Vital Knowledge founder.
But even though AI is driving enormous revenue growth for a wide swath of firms, it is only generating earnings/cash for a handful of companies, especially memory/chip companies, he notes.
South Korea and Taiwan markets are particularly popular and have recently surged past several long-established Western countries. They also house some of the most important companies in the entire AI ecosystem.
SK Hynix, a key supplier to AI chip giant Nvidia and Samsung Electronics, together account for over 40% of South Korea’s benchmark Kospi, while Taiwan Semiconductor Manufacturing Company TSMC accounts for over 40% of Taiwan’s Taiex benchmark.
While enthusiasm around AI continues to drive inflows, Crisafulli cautions that the pace of the rally may be difficult to sustain.
“It does seem like it’s in unsustainable territory,” he said. “You’ve got major stocks like Dell double in a handful of days, so the parabolic price action usually can’t last forever”.
The markets are going to have to absorb a lot of supply in the coming months so that’s going to be a big test, he adds.
The rally comes at a time when major tech companies are ramping up spending on AI infrastructure. Alphabet, Microsoft, Meta and Amazon are expected to pour over $700 billion this year into capex. Wall Street analysts estimate AI spending to climb above $1 trillion in 2027.
Crisafulli says the current environment shares some similarities with the late 1990s tech boom. But unlike the dot-com era, some of the biggest AI companies, such as OpenAI, Anthropic and xAI, remain private, making market evaluations difficult to access.
