AI ETFs Top $65 Billion, and the Chip Stock Trade Is Quietly Losing Its Grip
AI ETFs have pushed past $65 billion in total assets, according to J.P. Morgan data cited in a report published Sunday by The Daily Upside.
That makes the category one of the fastest-growing thematic fund segments Wall Street has ever seen.
Assets have roughly doubled since the start of last year — a pace no other sector ETF cohort has matched.
The growth points to something structural. Institutional capital is changing how it accesses the artificial intelligence trade, stepping back from concentrated bets on individual chip stocks and moving toward diversified, rules-based wrappers.
What An AI ETF Actually Is
AI ETFs are exchange-traded funds that hold baskets of publicly listed companies judged to have significant revenue exposure to artificial intelligence.
They work like any equity ETF. A fund manager defines an index or selection methodology, typically screening for companies deriving a threshold percentage of sales from AI-related products or services, then buys shares in those names in weighted proportions.
Investors buy and sell shares in the fund on a stock exchange throughout the trading day, gaining diversified AI exposure without selecting individual stocks.
The category spans a wide range of approaches. Some funds, like Global X (BOTZ) and iShares (IRBO), hold broad baskets of semiconductor makers, cloud platforms, and software vendors.
Others take narrower cuts, focusing on AI inference hardware, data-center real estate investment trusts, or companies providing AI-adjacent services such as cybersecurity and data labeling. The J.P.
Morgan data does not disaggregate by strategy, but the $65 billion figure covers the full spectrum.
Also Read: Chainlink CCIP Moves Over $7B As Institutional Use Accelerates
From Niche Theme To Mainstream Allocation
The report attributes the growth to two converging forces. First, Big Tech earnings over the past four quarters have repeatedly confirmed that AI-related revenue is measurable and growing, giving institutional buyers a fundamental anchor they lacked in 2023 and 2024.
Second, a wave of corporate AI deployment announcements in early 2026 persuaded pension funds and endowments that the AI trade was not purely speculative.
The timing matters. AI ETF assets stood at roughly $20 billion at the start of 2025, according to the underlying J.P.
Morgan data. Reaching $65 billion in approximately 18 months implies net inflows plus price appreciation at a combined rate that eclipses the early years of clean-energy and cannabis ETFs, two prior thematic cohorts that attracted intense retail interest before stalling.
The AI cohort’s distinction is that institutional allocators, not retail investors, are driving the largest flows.
Also Read: Has Wall Street Lost Patience With Big Tech’s AI Spending?
Why Institutional Buyers Moved To Wrappers
The structural reason for the ETF wrapper’s appeal is risk management. A fund holding 50 to 80 AI-exposed companies limits the damage from any single company’s earnings miss or regulatory action.
When shares of a major AI chipmaker fell sharply on export-control news earlier this year, diversified AI ETFs absorbed the shock without the drawdown that concentrated single-stock holders faced.
There is also a compliance dimension. Many large pension funds and sovereign wealth funds operate under mandates that cap exposure to any individual security.
An AI ETF satisfies those constraints while maintaining sector-level conviction. This is why the $65 billion figure almost certainly understates the total institutional bet on AI, since separately managed accounts and index-fund tilts toward AI-heavy sectors are not captured in the ETF category count.
The J.P.
Morgan data does flag concentration risk within the ETFs themselves. The top five holdings in the largest AI ETFs by assets overlap significantly, with the same handful of semiconductor and hyperscaler names appearing at high weights across multiple funds.
An investor who holds three different AI ETFs may believe they are diversifying when they are largely replicating the same exposure three times.
The Spending Debate That Could Reshape Flows
The $65 billion milestone lands at a tense moment. Meta, Microsoft (MSFT), and Amazon (AMZN) are all set to report earnings this week, and investor appetite for AI capital expenditure is under scrutiny.
Analysts at several banks have warned that the gap between AI investment and measurable AI revenue at the application layer remains wide.
If this week’s earnings calls produce guidance cuts or softer-than-expected AI revenue disclosures, AI ETF flows could reverse quickly. The category is liquid by design, meaning institutional sellers can exit within a trading session in a way they cannot with private AI fund positions.
That liquidity is a feature in rising markets and a risk factor when sentiment turns.
The $65 billion figure also draws an implicit comparison to the early-2000s tech ETF wave, which attracted record inflows in 1999 and 2000 before the dot-com correction erased roughly half the assets under management within 18 months. The parallel is not exact.
AI ETFs hold profitable companies generating real revenue, not pre-revenue internet startups. But the speed of the asset accumulation, doubling in 18 months, is historically unusual and warrants attention from anyone treating the category as a low-risk position.
Read Next: OpenAI CEO Says the Singularity Is Here, and He Is Not Speaking Hypothetically
