I model dealer gamma daily. Right now all three big index ETFs are negative-gamma below their flip,

I run a modeled dealer-gamma surface across the big index ETFs and the largest single names every day, and Friday's close showed a split I do not see often, so I figured this crowd would find it interesting.

As of Friday July 17 (first snapshot after monthly OpEx):

Index ETFs, all negative gamma, all below their zero-gamma flip:
\- SPY: net dealer gamma about -$8.7B, spot 743.29 vs flip about 751.7
\- QQQ: about -$5.1B, spot 695.33 vs flip about 717.4
\- IWM: about -$1.6B, spot 294.04 vs flip about 305.7

The six biggest components, all positive gamma:
\- AAPL +$1.0B, META +$0.85B, MSFT +$0.65B, NVDA +$0.47B, AMZN +$0.40B, GOOGL +$0.18B

So the wrappers are set up to amplify a move (dealers short gamma, hedging is pro-cyclical) while the stocks that make up most of their weight are set up to damp one (dealers long gamma). The usual explanation: index put-hedging keeps dealers short gamma at the index level, single-name call flow keeps them long gamma in the components.

Method: I model this from the listed options chain under an assumed dealer-positioning convention, so it is an estimate, not measured positioning. I snapshot it daily and rank each name against its own recent history.

The honest caveats, because they matter here:
1. Friday was July monthly OpEx, and this is the first snapshot after it, so part of SPY's swing (it was around +$5B a week ago) is just expiring positive-gamma July positions rolling off, not a fresh surge in hedging.
2. I have only tracked this surface for about three weeks, so I cannot call any single reading a record.
3. IWM has been negative gamma every single day in that window and the Russell 2000 was basically flat on the week, so negative gamma is a condition, not a forecast. It shapes how a move gets transmitted if a catalyst hits, it does not force one, into a data-heavy stretch that ends at the July 29 Fed.

Curious whether people here read the post-OpEx negative index gamma as durable, or a one-print reset that reverts as August positioning builds.

data + full method: [https://kresmion.com/daily-brief/2026-07-20?ref=reddit](https://kresmion.com/daily-brief/2026-07-20?ref=reddit)

https://preview.redd.it/qk7t9u66pdeh1.png?width=1608&format=png&auto=webp&s=02b7fab0a2c208a6a9e68bee79209035ce54483f