AGIF-All.Gl.Artif.IntelligenceI (ZZRF.F) closed at 299.78 EUR on October 16, 2026, down 7.2% in five days and 8.8% below its October 5 peak. The fund fell well over three times as much as the S&P 500, and no single news item explains why. The balance of evidence for the next two to six weeks leans negative, with medium conviction, because the daily and weekly trends point the same way and the fund lags the market.
Why it matters
There is no headline that explains this drop. The recent news is background: Nuveen's takeover of Schroders, a securities class action against Bloom Energy, a private-credit purchase of Fed pandemic loans. None of it is tied to the fund's holdings in the material provided, so it would be wrong to pin the slide on one name. The move looks like the AI theme itself being sold, and Friday's volume, 59.6% above the 20-day average, says the selling was not thin.
The market explains only a small part of it. The S&P 500 lost 1.97% over five days and 1.11% on Friday, while ZZRF.F lost 7.24% and 0.94% on the same stretches. Most of the five-day damage therefore came from the basket, not from the tape. No sector benchmark is available for comparison, so the theme's weakness cannot be split further. Over one month the fund is down 2.75% while the S&P 500 is up 1.22%. Over 50 days it trails the index by 6.7%. Year to date the fund is up 8.3% against 11.9% for the index, so the theme has been earning less than the broad market even before this week.
The argument about the theme has two sides. The constructive side is about demand for AI exposure. Bloomberg reported that Bain Capital Ventures raised $1.6 billion for early-stage AI startups, more than its previous $1.4 billion fund. On September 11, Reuters quoted Voya's Sebastian Thomas saying a U.S. listing could widen Kioxia's global appeal, after its shares surged 456% this year. The cautionary side sits in the same story. Voya's $14 billion AI fund does not own Kioxia because of liquidity, and a 456% run is the kind of move that leaves a theme exposed to a pullback. Separately, The Wall Street Journal carried the Bloom Energy class-action notice, a reminder that crowded AI-linked names carry legal and sentiment risk. Whether that stock sits in this basket is not shown. The Reuters piece is neutral on the question, and so is the evidence.
Valuation does not apply in the usual way. The fund has no analyst price targets or consensus. The only yardstick is the price path: the shares sit 16.0% below the 52-week high of 356.92 EUR and 16.9% above the low of 256.41 EUR.
What the evidence says
The long-term thesis is weakened, not broken. The price is below its 40-week average and its 10-week average, so the weekly trend is down. The fund has also dropped under the 200-day average at 310.31 EUR and the 50-day at 318.60 EUR. The year-to-date gain of 8.3% shows the theme still has real support underneath. But the pattern is a rally that peaked, rolled over, and has now slipped under the averages that buyers defended earlier. That is the early-to-middle phase of a downtrend, not a finished one.
For the next two to six weeks the balance of evidence leans negative. Conviction is medium, 2 of 4 points. Two things carry the reading: the daily trend runs with the weekly downtrend, and the fund is lagging the S&P 500 by 6.7% over 50 days. Conviction is not higher because the other two points are missing, and because the fund is already stretched to the downside.
The strongest case against this reading is that the move may be exhausted for now. The 14-day RSI is 30.8, which is oversold. The fund fell 7.2% in a week, a pace that rarely continues in a straight line. It is also within a whisker of 300 EUR, a round number where sellers have shown up on rallies before. A sharp rebound into that zone would be normal after such a drop. It does not tip the balance, because a bounce to 300 or to the 306.51 swing zone would still leave the price under the 200-day average and inside a weekly downtrend. That is why conviction stays at medium and not lower.
Levels that matter
The structure holds below 300 EUR. In a weekly downtrend, that round number is where sellers have appeared on previous rallies, with the 306.51 swing zone just above it. The trend would end only above 310.31 EUR, the 200-day average. Beyond that sit the 50-day average at 318.60 EUR and the round 325 EUR level.
The breakout level is a daily close below 275 EUR, the round number that is the first meaningful support. From there the only target in the data is the 52-week low of 256.41 EUR, with the next round level at 250 EUR beyond it. A breakout would be wrong above a stop at 301.47 EUR, which is 300 EUR plus a quarter of the 5.89 ATR. The math gives 0.7 to 1 in reward to risk. That does not reach 3 to 1. It would take a target far below the 52-week low, or a stop much closer to 275, to get there, and neither is on the map today.
What would change the picture
A weekly close above 310.31 EUR, the 200-day average, would be the first sign that buyers are back in control and that the weekly downtrend is losing grip. The non-price fact most likely to do it is evidence of real money returning to the AI theme: clear inflows into AI funds, or a Kioxia ADS listing that rekindles appetite for the group's leaders and is followed by the basket's holdings catching a bid rather than fading.
What to watch
- The 300 EUR level: whether rallies keep stalling there, or a daily close above 306.51 EUR starts to build a case.
- The weekly close against 310.31 EUR, the 200-day average, the line the trend must clear.
- A daily close below 275 EUR, which opens the way toward the 52-week low at 256.41 EUR.
- The S&P 500: it fell 1.97% this week but is still up 1.22% over a month. A stable market with this fund still falling would confirm that the weakness is in the theme, not the tape.
- Kioxia's U.S. listing timetable and the next reports on AI fund-raising, which are the clearest gauges of appetite for the theme. No dated earnings or macro releases are in the data, and holdings or flow figures for ZZRF.F are not available.