Owning Themes Generative Artificial Intelligence ETF (NASDAQ:WISE) is a bet that you can outsmart a basket. The fund holds 48 generative AI stocks, charges about 0.35%, and runs around $31 million in assets
If you bought WISE for tighter exposure to the generative AI thesis than a broad tech ETF offers, the question is whether that focus actually delivered, and whether the package around it (tiny AUM, single-theme correlations) ends up mattering more.
WISE works as advertised on concept and badly on math.
What WISE is actually selling
The pitch is pure exposure. Instead of a diversified tech fund where AI sits next to legacy software and payments, WISE concentrates into a few dozen names tied to model training, inference, accelerators, and applications. The return engine is straightforward equity appreciation in those names, with NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) as the gravitational center. NVIDIA just reported Q1 FY27 revenue of $81.61 billion, up 85% year over year. Jensen Huang called the AI factory buildout “the largest infrastructure expansion in human history”.
The problem is the rest of the bus. Alphabet (NASDAQ:GOOGL) is up about 124% over the past year on a Gemini and Cloud story that drove Q1 revenue of $109.9 billion and a backlog above $460 billion. C3.ai (NYSE:AI) is down 61% over the same year. Q3 revenue collapsed 46% year over year to $53.26 million, and GAAP gross margin fell to 17% from 59%. A 31-stock portfolio gives one weak holding meaningful drag. WISE owns both the engine and the flat tire.
Does the concentrated bet beat the broader bet?
Over the past year, WISE returned about 21%. The iShares A.I. Innovation and Tech Active ETF (NYSEARCA:BAI), an actively managed AI fund having exposure to bigger AI names, returned about 82% over the same period. Year to date the gap is about 41% for BAI versus about 0.5% for WISE. The performance differential is the central question every WISE holder needs to answer, because it tells you whether the concentration premise is paying off or quietly costing you the bull case it was supposed to capture.
The investor who picked the more concentrated generative AI bet did worse by a wide margin than the investor who paid 0.73% for a broader active sleeve covering the same theme. Concentration cut both ways. It avoided the chip-equipment names that ripped, and it loaded a struggling C3.ai into a 48-name portfolio where weak holdings are not diluted away.
The tradeoffs no one quotes from the fact sheet
- Single-theme correlation. Cannabis ETFs in 2019, cloud-computing baskets in 2022, and blockchain funds in 2018 all behaved like one position during drawdowns even though they held dozens of names. The diversification is cosmetic when the catalyst is shared. A training-data lawsuit, an AI capex revision, or a regulatory shock hits all holdings on the same day.
- Closure risk. WISE runs at about $31 million in AUM, below the $50 million threshold most issuers use to justify keeping a fund alive. ETFs that close typically give 30 days notice, and taxable holders eat the gain whether they wanted to sell or not.
- Liquidity at the wrap level. A small-AUM thematic ETF inherits the spread of its least-liquid name. In a stress tape, the bid you see on WISE is a function of the worst bid inside the 48, not the average, and that gap widens precisely when you want to exit.
Who WISE actually fits
WISE makes sense as a small satellite for an investor who already owns NVIDIA and Alphabet directly and specifically wants exposure to second-derivative generative AI names that broader funds underweight. It does not make sense as a primary AI allocation, because BAI and similar diversified vehicles have delivered materially better returns with lower closure risk and only modestly higher fees.
If the appeal of WISE is concentration, the last twelve months are the cleanest test of that thesis you are likely to get, and concentration lost. The honest read is that the fund is a tactical tool, not a core position, and treating it otherwise has been an expensive mistake over a full market cycle of generative AI enthusiasm.
