Information TechnologySemiconductors & Semiconductor EquipmentSemiconductor Materials & Equipment
CROWD: RED
Institutions are heading for the exits: 22 funds closed out last quarter (57 active holders down to 46), right after this name was crowded. Our data says this unwind phase skews badly. Not the dip it looks like.
Active holders?Institutions that actively chose this stock. Index funds and mega-diversified platforms (filers holding 1500+ names) are stripped out; they hold it automatically, not by decision.
46
of 123 total, index funds excluded
This quarter?Turnover of the active crowd vs the prior quarter: funds that opened a position vs funds that closed out entirely. The split shows texture: broad new interest vs a revolving door.
+12
22 out
Conviction?How big this position is inside the typical holder's own portfolio (median, in basis points). 1 bps is pocket lint they will not miss; 200+ bps is a bet they did real work on. Many holders at tiny weight is weaker than few holders at real weight.
1 bps
median holder weight: token bet
Streak?How many quarters in a row the active holder count has risen. Persistence, not a buy signal: long streaks are how names end up crowded.
0q
consecutive rising quarters
The crowd, 2016 to now
active holders per quarter
2016Mar 2026
Holdings as of Mar 31, 2026, filed by May 16, 2026. Counts for the newest quarter keep settling for a few weeks as amendments land.
The tide
funds in (up) vs out (down)
Top holders
by reported value · $585M held
Fund
Value
Note
CONTINENTAL GENERAL INSURAN…
$141M
Invesco Ltd.
$43M
Polunin Capital Partners Ltd
$36M
FRANKLIN RESOURCES INC
$35M
MACKENZIE FINANCIAL CORP
$29M
VANGUARD CAPITAL MANAGEMENT…
$29M
BlackRock, Inc.
$27M
MORGAN STANLEY
$22M
TWO SIGMA INVESTMENTS, LP
$22M
ARROWSTREET CAPITAL, LIMITE…
$17M
Tourists vs true believers
every holder a dot · right = bigger bet in their own book, up = patient money
Hover any dot for the fund and its bet size. Patience is measured from each filer's actual quarter-to-quarter turnover across our whole database.