ETF & Index Data Insights, News & Analysis | Ultumus

New Listings: A Language Model Now Picks Your Themes

Written by Bernie Thurston | Aug 27, 2026, 7:54:36 AM

The thematic fund industry has spent years asking investors to pick a future. It has now concluded, reasonably enough, that nobody can, and has outsourced the job to a machine.

 

iShares World Thematic Rotation Active UCITS ETF (THRW, London Stock Exchange)

This is an actively managed global equity fund whose entire purpose is to decide which theme you should be in, and then to change its mind. It holds somewhere between two and three hundred stocks, keeps the MSCI World index as a constraint, and charges sixty basis points for the privilege of rotating between short-, medium- and long-term themes as leadership moves.

The mechanism is the interesting part. The theme selection runs off a proprietary systematic model that uses data, artificial intelligence and large language models to rank and re-rank which stories are currently worth owning. A large language model, in other words, is now employed to have opinions about narratives. Given that themes are essentially narratives with tickers attached, this is either a category error or the most honest product design in the entire sector.

I lean towards honest. Every issuer that launched a single-theme fund made an implicit promise to be right about one thing for a very long time. Quite a lot of those funds have since been escorted out of the data quietly and without ceremony. A fund that promises only to keep changing its mind is, at minimum, harder to falsify.


Global X S&P 500 Quarterly Buffer and Quarterly Tail Hedge UCITS ETFs (SPQB and SPQH, London Stock Exchange)

Defined-outcome products continue their steady march across European currency lines, and these two are the tidier end of the risk-engineering business. The buffer version absorbs the first five per cent of quarterly losses on the S&P 500 and hands you the upside up to a cap. The tail hedge version leaves the first three per cent of losses with you, then covers the next nine, on the theory that small losses are survivable and large ones are not.

Both charge fifty basis points. Both are constructed on published Cboe indices rather than a black box. There is nothing absurd here at all, which in this batch counts as a distinguishing feature.

Which makes the next entry rather awkward.


iShares Russell 1000 Value UCITS ETF (R1VG, London Stock Exchange)

While the rest of the queue rotates, hedges and re-ranks, one fund has turned up in London offering European investors large-cap American companies that are cheap relative to their book value, and nothing else. The benchmark is capped to keep any single name from dominating. There is no volatility target, no theme, no buffer and no model deciding what the story is this quarter.

Value has spent long enough out of fashion that a plain value index fund arriving in a new market now reads as a contrarian statement rather than a shelf-filling exercise. It probably is not one. It is probably a distribution decision. But I enjoyed it anyway.

iShares USD Emerging Markets Bond UCITS ETF, GBP Hedged Distributing (HCGH, London Stock Exchange)

And then, at the bottom of the batch, this. A sterling-hedged distributing share class of an existing emerging market bond fund. Somebody in the UK wanted the income in pounds without the dollar risk, said so, and somebody else built the share class.

No leverage. No rotation. No language model. No volatility target. Just a currency hedge and a distribution schedule, quietly doing the single job it was asked to do.