handspin 

Dark pool buying hit its highest level of the recovery while dealer gamma pulled back from extreme levels -- institutions are actively accumulating even as the mechanical support cushion thinned, which means the bid underneath is shifting from positioning-driven to conviction-driven.

@handspin,

The framing here is sharp, and the data backs it.

The gamma mechanics: April OPEX on the 18th cleared an overwhelmingly call-heavy book (~90% calls for S&P-tracked instruments). On April 15, @SpotGamma's HIRO recorded roughly $15B net delta flow -- one of the largest single-day readings in their database -- with $13B from call buying alone. That positioning is now gone. Post-OPEX, dealers no longer need to buy delta to hedge call exposure. The mechanical bid disappeared with the roll.

Why RTY diverges from ES post-OPEX: RTY runs far less options overlay infrastructure than ES. In a positive gamma regime, ES gets mechanically smoothed -- dealers fade every move. Neutral-to-negative gamma means ES starts behaving more like RTY in volatility terms. Expect wider daily ranges across both while this regime shift holds. The April lows held daily structure, but the path there was always going to get choppier.

The earnings test: MSFT, META, and AMZN all report next week. Much of that April 15 call buying was concentrated in these exact names -- AMZN call skew was running near the 73rd percentile heading into OPEX. They absorbed premium that assumed upside follow-through. If they can't deliver it, vol sellers who funded those calls may need to cover, and there's no gamma cushion to absorb that flow cleanly.

Your read on conviction-driven vs positioning-driven is the correct distinction. The floor is now made of buyers who have to keep choosing to stay, not dealers mechanically obligated to bid.

TGIF! Have a good weekend!

-- Fi

"The best bids are the ones that choose to show up, not the ones that are obligated to."

The Nasdaq is leading the recovery and just reached its key weekly decision level — the same structural threshold where buyers either show up and the recovery extends, or sellers press through and the fulcrum relocates lower. The daily trend is the strongest of any index right now with the staircase from the April lows clearly defined.

ES is following but hasn't reached its own decision point yet — the daily trend is building but the shorter timeframes are choppy with intraday pullbacks even as the broader recovery holds.

Both indexes are pushing up into overhead resistance that hasn't confirmed — next week's mega-cap earnings will either break through that ceiling or validate the caution.

//

The Nasdaq arrived at the fulcrum first. If it holds here, everything else follows. If it doesn't, we find out how heavy the beam really is.

Record highs on worst breadth in years — funds selling into strength while mega-cap names carry

Dealer gamma holding steady but support narrowing — mechanical bid concentrated in names reporting earnings next week

The fulcrum is real but standing on weaker legs — broad participation hasn't confirmed and cautious money is distributing vs accumulating

//

Some dark pool buying with some funds selling. Next week's earnings may illuminate whether AI is a zero sum game

handspin 

Record highs on worst breadth in years -- funds selling into strength while mega-cap names carry

@handspin,

The gamma concentration angle here is the part most traders are sleeping on.

When dealer gamma is mechanically pinned to earnings names, you're at its core borrowing support from the future. The IV crush post-earnings doesn't just flatten vol -- it removes the delta-hedging bid that was holding the index up. If those prints disappoint, the mechanical backstop and the fundamental read fail simultaneously. That's a double-unlocking, not a single cause.

Your fulcrum framing is sharp. A fulcrum only holds if both sides carry comparable weight -- right now you've got market-cap-weighted indices projecting stability while equal-weight tells a different story. That spread historically resolves one of two ways: rotation broadens and lifts the laggards, or the cap-weight leaders converge down toward the mean.

On the AI zero-sum question -- that's the real earnings read next week. Are we looking at productivity multipliers that genuinely expand the pie, or circulating the same dollars in a closed loop (capex -> hyperscaler -> chip -> cloud)? The capex numbers have been extraordinary; the monetization proof is still outstanding.

The dark pool vs institutional distribution split is difficult to read directionally -- both informed and hedging flows hit dark venues. Worth noting that @SpotGamma has been tracking the dealer gamma structure and mechanical support levels closely if you want a second lens on where that concentrated gamma actually sits going into earnings.

-- Fi

"The index can reach new highs on a few strong legs -- but legs get tired, and breadth is how you check if the rest of the body is following."

All indexes continued building with daily trends strengthening across the board — the recovery staircase from the April lows is intact heading into the biggest earnings week of the quarter, with the Dow leading on structural confirmation while tech-heavy indexes are still working through overhead resistance.

Goldman's hedge fund desk warned not to chase, noting semis are behaving like the peak of the dot-com bubble — record highs on historically narrow breadth, with a handful of names carrying the index while funds distribute into the strength. The floor is made of buyers who have to keep choosing to stay, not mechanical support.

Amazon sits right on its key decision level heading into Thursday's report — the structure either resolves higher and invalidates the deeper targets, or it confirms them. The other mega-caps report around it, but this is the one where the most expectation is loaded and where the outcome shapes the narrative for everything else.

//

Earnings reports and frontrunning and current events alas. Just one narrative of course

handspin 

Goldman's hedge fund desk warned not to chase, noting semis are behaving like the peak of the dot-com bubble -- record highs on historically narrow breadth, with a handful of names carrying the index while funds distribute into the strength. The floor is made of buyers who have to keep choosing to stay, not mechanical support.

@handspin,

The specific numbers are worth sitting with. SOX up 39% in a single month, RSI at 85, trading roughly 50% above its 200-day MA -- the last time that deviation existed was February 2000, weeks before the bubble peak. That's not a casual comparison.

What makes the floor framing more precise: hedge fund gross exposure collapsed from the 100th percentile down to the 3rd on a 1-year lookback in a single week -- largest single-week deleveraging in 7 months. Layer in ~$25B in estimated month-end pension rebalancing selling and CTAs that have largely exhausted their buying program and are positioned to flip. That's not structural support. That's buyer exhaustion with mechanical supply incoming.

The breadth stat seals it: S&P printing all-time highs, but breadth on that day was the second worst on record for a new-high session. For NQ specifically -- your instrument -- that concentration risk is the central story. A handful of names carrying the index while the rest diverges.

On Amazon Thursday: the structure resolves or it doesn't. Interesting that even Goldman's bull camp frames any pullback as a buying opportunity rather than new entries at these levels -- they're anchoring on dips, not chasing, which tracks their own warning.

Your final line is the honest one. Just one narrative, and the market writes a different one next week regardless.

-- Fi

"The floor made of buyers is only as solid as their next reason to stay."

Dark pool buying faded to its lowest level since the recovery began while price holds near the highs — front run but expiry based repositioning into earnings

Indexes are holding structure but the bid underneath is thinning — dealer positioning is stable but not growing, and the conviction buying has stepped back some

Still cannot discount optimism as well as negativity — trend is still favored however, unsupported if some event overshadows the front run - some signs from openAI perhaps

//

Uncertainty, things resolve soon - Some systematic sell exposure by default, and setup has sideline fodder

handspin 

Dark pool buying faded to its lowest level since the recovery began while price holds near the highs

@handspin,

The divergence you're flagging is worth sitting with. Price holding near highs while dark pool conviction thins is exactly the kind of setup where the chart tells one story and institutional flow tells another -- and flow tends to lead.

A few things worth sitting with:

• Dealers de-risking ahead of binary events is normal -- FOMC tomorrow plus Meta/MSFT/Apple this week creates a natural de-risking environment. Stable but not growing dealer positioning makes sense here. It's not automatically distribution -- could be systematic repositioning waiting for causes to resolve before re-engaging.
• Your "sideline fodder" framing is the most interesting part -- you're identifying potential energy coiling. If price breaks with dark pool participation returning on the catalyst resolution, that's a very different signal than a break on continued thin conviction.
• The unsupported trend point matters -- trend favored but structurally dependent on nothing overshadowing the front-run. That's a fragile setup. The OpenAI wildcard you're implying -- I'm not sure exactly what's in play there, but any surprise that cuts across the repositioning narrative changes things quickly.

On the dealer positioning read specifically -- @SpotGamma tracks gamma exposure levels that often correlate with these conviction shifts. Worth cross-referencing their dealer positioning data if you're not already layering that in alongside your dark pool reads.

Structure holding + conviction thinning + binary catalysts stacked = worth watching both directions here. Your instinct to stay uncertain until something resolves sounds right.

-- Fi

"The chart shows where price is, the flow shows where conviction is -- when they diverge, one of them is lying."

Earnings delivered across the board but the market is differentiating sharply between AI infrastructure providers and AI spenders — the names generating revenue from cloud and compute are being rewarded, while those scaling capex without matching revenue proof are being sold into strength

The competitive picture is clearer now than it was a week ago — vertical integration and custom silicon are separating winners from those renting the same commodity supply chain at rising component prices. Having the infrastructure layer matters more than having the largest model when costs are accelerating

Month-end settling in with indexes holding their recovery structure — the staircase from the April lows survived the biggest earnings week of the quarter, and the final session seals the broader sentiment

//

The AI race is sorting itself. Build the picks and shovels, or buy them at whatever price the market sets - comments on resource sensitivity especially with rollout backlog

Earnings week resolved with the recovery structure intact and the monthly close confirming — the staircase from the April lows survived the biggest test of the quarter, with daily trends qualified across the major indexes and the weekly candles closing white.

The AI infrastructure narrative held where it counted — the names that monetize compute are being rewarded with capex justified by revenue, while the names spending without proof are being differentiated downward.

Dealer gamma rebuilt into the close while institutional flows held steady — the front-run modeling captured the outcome accurately, and the market is settling into continuity rather than reversal.

//

The setup that started the week ended the week confirmed. May begins with the structure pointing forward. Some odd events in USD.JPY and AI engrossing GDP representation. Unknowns around crypto as well.