The same fund names obligation that reaches a sector fund reaches a thematic ETF, and it works the same way once a name suggests a focus. Under 17 CFR 270.35d-1, a fund whose name indicates that its investment decisions incorporate a particular focus must adopt a policy to invest at least 80% of its assets in accordance with that focus. A fund named for artificial intelligence or clean energy is making that kind of claim, so it is bound by the same 80% requirement as a fund named for an industry. The mechanics of maintaining and monitoring that basket are the same across every fund the rule reaches, and are covered on the page for growth stocks. What differs for a thematic fund is not the obligation but who decides what the 80% is measured against: for a sector fund, an established industry classification does that work; for a thematic fund, the provider's own written methodology does.
That difference in who writes the definition is the central fact about this category. Because no regulator or standard-setting body defines "what counts as an artificial intelligence company," each provider's index methodology sets its own inclusion rules, and those rules commonly involve screening company disclosures, revenue segments, or patent and research data for exposure to the theme, then applying a scoring or ranking process the provider designed. Two ETFs built around the same theme can therefore hold materially different companies, in different proportions, while both are legitimately built around it. Reading the fund's stated methodology, not just its name, is the only way to know what a given thematic ETF actually owns.
The cross-sector construction has a real diversification consequence that is easy to miss. A thematic ETF can look diversified because it holds companies nominally drawn from several conventional sectors, and that appearance is misleading. Every holding shares one thing: exposure to the same theme. If the market re-prices that theme, whether because the narrative cools, a competing technology emerges, or the underlying trend simply takes longer to play out than expected, most of the fund's holdings are exposed to the same repricing at once, in a way that companies drawn from unrelated sectors for unrelated reasons would not be.
Costs run higher, and the reason is structural rather than incidental. Building and maintaining a thematic index requires ongoing research judgment about which companies still qualify, which is more labor-intensive than tracking an established broad or sector index, so thematic ETFs typically charge higher expense ratios than either. That expense compounds every year the fund is held, regardless of how the theme performs.
Performance-chasing is the hazard specific to this category, and it is documented rather than merely anecdotal. New thematic ETFs are frequently launched, and existing ones frequently attract new assets, after a period of strong performance in the underlying theme, since that is when investor interest is highest and a launch is commercially viable. A fund built and marketed around a theme that has already run up is disproportionately likely to be bought near a high point in that theme's popularity, which is the opposite of the entry timing an investor would want. This dynamic does not make every thematic ETF a bad investment; it means the fund's recent return is a poor guide to whether now is a good time to buy it.