📊 Ainvest Option Flow Digest — 2026-07-30

The most hedged tape in weeks: ≈$294M printed across 10 names, but almost all of it came bundled with a stock hedge, was a roll, or was a position being closed — only ≈$15M was a real directional bet

✅ RESOLVED — July 31 Pre-Market OI Check on This Session's Flags

All 16 provisional ⏳ legs from this session are now settled by the July 31 OPRA open-interest snapshot: 14 confirmed OPEN, 2 confirmed CLOSE, and no directional read inverted. Two things changed in the numbers below, and both are corrected throughout this digest:

• 🔄 SMH was smaller than reported — ≈$44M, not $49M. One of the two 5,000-lot call clips was cancelled by the exchange and re-reported at 2,000, so the call sale is 8,000 contracts for ≈$17.6M, not 10,000 for $22M. The open-interest build (+8,139) independently confirms the corrected size. The hedged, short-volatility read is unchanged — only the scale.
• ✅ LRCX's $45M put is a CONFIRMED close, not a guess. Open interest fell 16,783 → 9,344 (−7,439) against a 7,500-lot print — a 99.2% match. Nobody put on $45M of new bearish risk; somebody retired an old hedge. The fresh-short scenario is ruled out.

Everything else confirmed as published: MSFT opened bigger than its own print (OI 10,736 → 21,511, +10,775 vs 9,953 — the bullish read is proven, not provisional); FLY built ≈164% of its print (143 → 5,762, other buyers joined); EWY's protective put was not the roll we flagged as possible — it is brand-new insurance (4,553 → 11,768); CORZ opened in full (3,105 → 8,163); AVTR matched to 0.3% (138 → 23,190); EQIX's ambiguous near leg came back a genuine fresh buy (3,181 → 6,022); AMGN's roll is confirmed on both sides (August closed −1,726, October opened +2,194); and all four QQQ legs opened — including the Sep $710 call sale (29,510 → 40,478), which proves the calendar was built rather than unwound. One asterisk there: QQQ's June-2027 $600 put built only ≈41% new contracts, so that leg added less net exposure than its size suggests.

The standing lesson, twice over today: a headline premium is provisional until the next-day open interest confirms it — and it can be wrong in size as well as in direction. A cancelled clip took 25% off SMH's call ticket, and only the OI cross-check caught it.

🎯 Quick Read

If you only remember one thing today: a giant premium number that arrives with a matching block of stock is not a directional bet. Almost every big ticket today was delta-hedged at the instant it printed.

• ≈$294M gross printed across 10 names (revised down from ≈$298M after a cancelled SMH clip — see the ✅ box above), but only ≈$15M is genuinely directional — two lit, at-the-ask bullish call buys (MSFT $13M, FLY $2M). Everything else was hedged, a roll, or a position being closed the moment it traded — including the day's late tickets: a $45M LRCX ITM put now confirmed as a hedge being closed after the stock jumped ≈17%, an AMGN dividend-timing call roll, an EQIX put calendar, and an AVTR call overwrite after a ≈14% pop.
• The $122M QQQ "call buy" is NOT bullish — it printed with an ≈825,000-share stock hedge, making it a delta-neutral volatility/convexity structure on the Nasdaq-100 (plus call and put calendars), not an upside bet.
• The ≈$44M SMH premium was SOLD and hedged — a covered call + a delta-hedged short LEAP put = a range-bound, short-volatility harvest on semis, not a direction.
• The Korea and Core Scientific puts are protection, not doom bets — EWY ($12.8M) and CORZ ($2M) both came bundled with matching long stock, i.e. someone hedging a position they own.
• MSFT is the one clean conviction bet: a lit buyer pressed $470 calls the morning after Microsoft's blowout (Azure +43%, past $100B annual) — real, unhedged, directional.

The lesson before you copy any of this: five of six tickets today were desks hedging, harvesting premium, or protecting — the opposite of "load up." Read the stock hedge, not the headline dollar figure.

🔁 OI Review — Last Session's Provisional Flags, Now Resolved

The July 30 pre-market open-interest snapshot is in, and Wednesday's tape held up far better than Tuesday's. Nine tickers carried ⏳ provisional flags across 12 legs. Eleven resolved exactly as published — several bigger than we predicted. One inverted: Eaton's near-dated put leg.

🔄 The one inversion:

• 🐻 ETN — the "$6.2M put diagonal" was a bearish ROLL: the Sep $310 put opened cleanly (+5,718), but the Jul-31 $350 leg's OI fell −579 — a print that retires open interest is an exit, not an open. The desk moved its bearish position down in strike and out in time, past Eaton's earnings. Genuinely new money ≈$3.5M, not $6.2M — same direction, de-risked runway.

✅ Confirmations — three came in above target:

• 🐻 VFC OI 10,209 → 50,210 (+40,001) — ≈99.9% of session volume; our buyer was only ≈65% of it, ≈14,000 more opened by other desks.
• 🐻 NBIS OI 3,410 → 62,043 (+58,633), an ≈18× build — the 2-day crash hedge was a crowd.
• 🐂 INFY OI 1,146 → 109,038 (+107,892), ≈95× (a separate 10,000-lot cross was busted and left no imprint).
• 🔵 BHP +30,060 (≈100.2%) · 🔵 GEV +2,535 (≈102%) · 🛡️ BE both collar legs opened · ⚪ GLW $115 call opened (+15,067), $310 put closed to 80 · ⚪ META +15,994 (≈100%) — combined with 7/28, a two-day ≈27,600-contract programmatic overwrite campaign.

The standing lesson, with a twist: a big BUY headline is not conviction — only next-day OI tells you if money opened, closed, rolled, or transferred (Eaton). But this session also showed the mirror: three reads resolved bigger than published. Next-day OI is not only a bust-detector — it reveals when the bet you saw was the smaller half of a crowd.

📋 At a Glance

🗓️ Weekly · 📅 Monthly · 📊 Quarterly · 🏛️ LEAP (>1 year)

🔎 The Ten, in Plain English

🐂 MSFT — Microsoft: the day's one clean conviction bet. The morning after Microsoft's blowout fiscal-Q4 — revenue $90B (+18%), Azure +43% and past $100B in annual revenue, stock up ≈9% after hours — a buyer lifted the offer on ≈9,953 September $470 calls (≈$13M). This is lit, at-the-ask, unhedged buying: someone pressing the post-earnings move rather than fading it. The $470 strike is ≈5% above spot with ≈7 weeks to run. Real directional bullishness — the cleanest read on the board.

🐂 FLY — Firefly Aerospace: a small, convex recovery bet on a beaten-down space name. Firefly IPO'd at $45 in 2025 and sits near $19, down ≈70%. A buyer paid the offer for ≈3,420 February-2027 $20 calls (≈$2M) on a strike that held almost no open interest — a genuine fresh position. The lure is Blue Ghost Mission 2, a far-side lunar landing slated for late 2026, inside the option's life. It's real bullish demand, but on a volatile, loss-making small-cap with share overhang — a high-risk flyer, not a core holding.

⚪ QQQ — Nasdaq-100 ETF: $122M that looks like a moonshot and isn't. The headline is ≈$75M buying March-2027 $715 calls — but ≈825,000 shares of QQQ printed alongside as a hedge, zeroing out the directional exposure. Add a sold September $710 call (a call calendar) and a $600 put calendar, and what you have is a sophisticated, delta-neutral volatility and term-structure position on the Nasdaq-100 — placed into a brutal mega-cap earnings week (Microsoft +14%, Meta −9% on 7/29; Apple and Amazon tonight) and the day after the Fed held. It is not a bet on which way the index goes. A $122M number that says "positioning," not "conviction."

⚪ SMH — Semiconductor ETF: ≈$44M collected to bet chips go quiet, not up or down. Two premium sales, both hedged with stock: 8,000 August $550 calls sold against a matching stock block (a covered call — note one 5,000-lot clip was cancelled and re-reported at 2,000, which is why this is $17.6M and not the $22M first reported), and 3,000 January-2028 $500 puts sold against an ≈84,000-share hedge (a delta-neutral short put). Both legs are confirmed opening by the next-day OI. Net: a desk harvesting ≈$44M of elevated volatility on a range-bound $500–$550 view after the SOX's ≈20% July drawdown — betting semis chop sideways into Nvidia's August 26 print, not that they break out or crash.

🛡️ EWY — South Korea ETF: protection, bought after the fire started. An institution bought ≈7,500 October $150 puts (≈$12.8M) — but ≈300,000 EWY shares printed alongside, the signature of a protective put on a position they own, not a naked short. The context is dramatic: the KOSPI ran 100%+ to a June record, then crashed ≈33% in July after China began mass-producing homegrown chipmaking tools — a direct threat to the Samsung/SK Hynix memory names that are ≈44% of the fund. Even after the plunge, Korea is still up sharply on the year, so hedging the rest of the gains is rational. (Note: the strike's open interest was slipping, so this may be a roll of an existing hedge — tomorrow's OI will tell.)

🛡️ CORZ — Core Scientific: a cheap seatbelt on an AI-datacenter transformation. Core Scientific just landed a $14B, 15-year AMD data-center deal (7/28) that pushed its backlog past $24B — but it came with a dilutive share warrant and a ≈$6B build-out whose revenue is weighted to 2027-2028. So a desk bought ≈5,000 deep-out-of-the-money December-2027 $15 puts (≈$2M) — hedged with ≈95,000 shares — as inexpensive insurance against 12-18 months of financing and execution risk. Downside protection on a story they still believe in, not a bearish call.

⚪ LRCX — Lam Research: a $45M put "sale" that is confirmed to be a hedge coming off. On the day Lam jumped ≈17% on a blowout quarter (September guide ≈$1B above the Street), a desk sold 7,500 September $340 puts — deep in-the-money, ≈$45M, mostly intrinsic value. The tell was the open interest, which was falling — and the next-day snapshot settled it: OI fell 16,783 → 9,344 (−7,439) against a 7,500-lot print, a 99.2% match. A sale that shrinks open interest is a close, not a fresh bet. Someone who owned those puts as downside protection let them go now that the stock ripped — removing a hedge, not making a bearish call. Neutral-to-bullish position management, and a reminder that a big "sell puts" headline can mean the opposite of what it looks like.

⚪ AMGN — Amgen: housekeeping, not a bet. A deep-in-the-money $310 call position (≈$72 above the strike) was rolled from August to October — same strike, further out in time, for a tiny ≈$0.4M net cost. Why? Amgen goes ex-dividend August 24, and the August calls expired three days before that; rolling out dodges early-assignment/dividend headaches while keeping the long exposure alive. This is calendar mechanics — financing and dividend timing — not a directional view on the stock.

⚪ EQIX — Equinix: a tactical fade toward $1,000. The day after Equinix beat but guided Q3 light (≈−3.5%), a desk put on a $1,000 put calendar — buying the near August put and selling the more expensive November one at the same strike, for a net credit. The near leg doesn't span an earnings event, so this is a short-term drift-and-vol play toward the round $1,000 number, not a crash bet. Neutral, tactical positioning.

🔵 AVTR — Avantor: selling the pop. After Avantor jumped ≈14% on a Q2 beat and its first lab-solutions growth in quarters, a seller wrote ≈23,000 September $14 calls (≈$1.7M collected) right at the top of its low-teens analyst range. It's an overwrite — capping upside near the ≈$14 fair-value ceiling and monetizing the earnings volatility spike. Short-upside income, not a crash call — though it does say a pro thinks the easy money post-pop is made.

👥 How Four Different Traders Might Read Today

🎰 YOLO / momentum trader — The only clean directional prints are MSFT $470 calls (bullish, but chasing the day-after-earnings pop is exactly when you overpay for premium) and FLY $20 calls (a genuine lottery on a Moon-landing catalyst — tiny position size only). The $122M QQQ and ≈$44M SMH tickets are hedged — there is nothing to chase there.

📈 Swing trader — The pros are telling you their ranges: SMH sellers see $500–$550 into Nvidia (Aug 26); the QQQ calendars are positioned for a volatility move, not a direction, into the AAPL/AMZN prints tonight. Respect those levels. MSFT's post-earnings strength is real, but wait for a pullback toward gamma support rather than buying the gap.

🏦 Premium collector — This was your tape. SMH is a textbook hedged premium harvest — covered calls plus a delta-neutral short LEAP put, ≈$44M collected on a range-bound view. The lesson to copy is the hedge, not just the sale: every short leg today came with stock next to it. Naked premium selling into an event is how accounts blow up.

🌱 Beginner — Today's one big idea: look for the stock hedge. When a $122M "call buy" prints with ≈825,000 shares beside it, the trade has no view on direction — it's a bet on volatility. Most of today's biggest numbers were like that. "$122M" and "$44M" tell you how much size, not which way. Learn to ask "was it hedged?" before you ask "is it bullish?"

📆 Upcoming Catalysts (catalyst ≠ option expiration — keep them separate)

⚠️ Risk Control & Patience

Unusual options flow is a research starting point, not a signal to copy — and today is the clearest example yet. The three biggest dollar figures on the board (QQQ, SMH, EWY) were all hedged or protective: a delta-neutral volatility structure, a premium harvest, and portfolio insurance. None of them is a directional bet, and copying the option leg without the stock leg would leave you with the opposite exposure the institution actually wanted.

Before you act: (1) look for the paired stock block — a matching hedge means the trade has no directional view; (2) confirm open-vs-close with the next day's open interest — today's flags have now all resolved (see the ✅ box up top), and that check caught both a crowd joining MSFT and FLY and a cancelled clip that shrank SMH by 25%; (3) size so a single earnings gap can't hurt you; (4) never sell premium naked into an event. The institutions today were managing risk. That is the part worth copying.

Not financial advice. Options involve substantial risk. Institutional flow reflects one side of a trade whose full intent, hedges, and counterparty we cannot see.

Last updated: 2026-07-31 — the July 31 pre-market OPRA open-interest snapshot resolved all 16 provisional legs from this session (14 OPEN, 2 CLOSE, no directional inversions). SMH's call-leg size and all its premium figures were corrected downward after an exchange cancellation was found on the tape (≈$44M program total, was $49M).