📊 Ainvest Option Flow Digest — 2026-07-29

Wall Street sold ≈$44M of downside insurance on the power names — while two earnings losers drew real bearish buyers

🎯 Quick Read

Today was earnings-driven and hedge-heavy — the biggest dollar tickets carry the least directional signal, so read the structure, not the headline number.

• ≈$173M gross printed across 9 names, but only ≈$15.4M is genuinely new directional money — mostly bearish (VFC + ETN + NBIS buying puts), plus one speculative bullish flyer (INFY calls). (Revised down from ≈$18M on July 30: next-day open interest proved ETN's near-dated leg was a roll-close, so only its ≈$3.5M September leg is new — see the July 29 Resolution block below.)
• ≈$82.6M of upside/downside insurance was SOLD, not bought — three desks collected premium: META $39M (covered-call overwrite into tonight's earnings), and the power names GEV $30.1M + BHP $13.5M betting their floors hold. Income structures, not crash bets.
• ≈$48.7M in Corning looks huge but isn't directional — a delta-hedged convexity structure plus a closing trade. A big premium number ≠ conviction.
• The one costless collar (BE) is pure downside protection on a wild name.
• Two earnings losers, two different reactions: VF Corp fell ≈17% today and drew aggressive put buyers; Corning fell ≈17% yesterday and drew a hedged institution, not a dip-buyer.

The lesson before you copy any of this: the loudest premium prints today were desks selling insurance or hedging — not punters loading directional bets. Match the structure to your own risk before you follow.

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

The July 29 pre-market open-interest snapshot is in, and Tuesday's tape split cleanly down the middle. Eleven tickers carried ⏳ provisional flags into this morning. Seven resolved as published — including the ≈$125M SanDisk headline, which cleared its bust check. Four inverted: two prints turned out to be pure transfers that created no position at all, and two were outright closes that meant the opposite of their "BUY" label. Net effect: the session's genuinely-new premium drops from ≈$305M gross to ≈$197M, and its directional new money to ≈$136M.

🔄 Inversions:

• NVDA — the "clearest bullish bet" was a transfer: the Dec-2027 $210 call OI fell 391 against a 4,606-lot print. ≈4,600 LEAP calls changed hands; nothing new was created.
• GLW — a "$7M long-convexity bet" was a covered-call being unwound (BTC): the Sep $140 call OI collapsed 10,862 → 2,457 (≈−82%), paired with a ≈357K-share sale.
• TLN — a "$10M bearish put buy" was somebody covering a losing short put (BTC): OI fell by more than the entire print.
• CRWD — a "$15M new overwrite" was an existing one changing hands (transfer): the $195 call OI slipped 250 on the strike's entire session volume.

✅ Confirmations: SNDK ≈$113M held (OI +2,275, bust check cleared) · RCL confirmed and bigger (OI +10,596) · INTC opened but still delta-hedged financing (OI +9,796) · META ≈73% opened covered-call overwrite.

The standing lesson: a big BUY headline is not conviction. Only next-day open interest resolves whether money actually opened a position — or just closed, rolled, or transferred one.

✅ July 29 Resolution — Added July 30, 2026

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

🔄 The one inversion:

• ETN — the "put diagonal" was a bearish roll, not a two-layer bet. The Sep $310 put opened cleanly (3,165 → 8,883, +5,718), but the Jul-31 $350 put's open interest fell (5,770 → 5,191, −579) — a print that retires open interest cannot be an open. That leg was the exit side: the desk released its near-dated $350 exposure on the eve of expiry and moved the position down in strike and out in time, past the earnings print. Only the ≈$3.5M September leg is new money; the true net debit may be under ≈$1M rather than ≈$6.2M. Still bearish, just re-priced rather than doubled.

✅ Confirmations — several came in above target:

• VFC OI 10,209 → 50,210 (+40,001), ≈99.9% of the strike's session volume and above our predicted 26,000–36,000. The cleanest bearish bet of the day got cleaner — and it had company.
• NBIS OI 3,410 → 62,043 (+58,633), an ≈18-fold build. Our print was only ≈52% of it; other buyers loaded the same 2-day crash hedge.
• INFY OI 1,146 → 109,038 (+107,892), a ≈95-fold build. A separate 10,000-lot cross on the strike was busted and left no imprint.
• BHP OI 55 → 30,115 (+30,060) — ≈100.2% of the print. Textbook.
• GEV OI 162 → 2,697 (+2,535) — ≈102%, above target. The day's biggest credit is real and new.
• BE both collar legs opened above target: calls 839 → 4,801 (+3,962), puts 73 → 4,068 (+3,995). One structure, both sides confirmed.
• GLW both reads confirmed — $115 call opened (219 → 15,286, +15,067) and the $310 put closed to 80 (−920) against our predicted ≈76-100. A bullseye on the harder call.
• META OI 229,279 → 245,273 (+15,994) — ≈99.96% of the print, a near-perfect fresh open. New finding: combined with July 28's confirmed +11,638 at the same strike, this is a two-day, ≈27,600-contract programmatic overwriting campaign, not a single opportunistic sale.

What changed: the session's genuinely-new directional money drops from ≈$18M to ≈$15.4M (ETN's near leg was an exit, not an entry). Gross premium (≈$173M) is unchanged — that money did change hands. The lesson cuts both ways this time: most reads held, and two of them were bigger than we published. Next-day open interest is not only a bust-detector — it is also the only thing that can tell you a bet was larger and more crowded than the single print you saw.

📋 At a Glance

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

🔎 The Nine, in Plain English

🐻 VFC — VF Corporation: the day's one clean directional bet. VF Corp reported a fiscal-Q1 loss this morning and fell ≈17% on the Vans-brand drag (Vans −9%) — even though revenue beat and the full-year revenue guide went up. Into that drop, a buyer pressed the offer three separate times to accumulate ≈26,000 March-2027 $14 puts (≈$4.6M) as the stock slid from $15.33 to $14.75. This is real lit buying — someone lifting the ask, not a negotiated block — and it's the clearest bearish read of the day. Note: these are ≈1.7-year puts (a patient position/hedge on a 2027 refinancing overhang), not a one-day gamble.

🐻 ETN — Eaton: a bearish position rolled past earnings, not doubled into them. A worked complex order printed two put legs at the same second — a Sep $310 put (≈15% out-of-the-money tail) and a $350 put expiring this Friday, the day Eaton reports Q2. Eaton ran ≈30% this year on the AI-datacenter electrification story, then de-rated into the print on valuation. 🔄 Resolved July 30: open interest proved this was a roll, not a diagonal — the Sep $310 leg opened (+5,718) while the Jul-31 $350 leg's OI fell (−579), meaning that leg was the exit. The desk traded a 2-day binary strike for ≈7 weeks of runway below $310. Genuinely new money: ≈$3.5M, not ≈$6.2M. Still downside positioning — but a considered de-risking inside a bearish view, not an escalation.

🛡️ BE — Bloom Energy: a two-year seatbelt, bought for ≈nothing. Bloom printed a record Q2 (7/28) on AI-power demand (Oracle's 2.45 GW Project Jupiter, Brookfield's $25B framework) — but has whipsawed ≈47% off its peak after a July short-report. So a desk built a wide collar: sell the far-away $350 call to pay for a cheap $110 crash put. Net cost ≈$0.2M — essentially free insurance. The ≈$24M is notional across both legs, not money spent. It's protection, not a directional view.

🔵 BHP — BHP Group: a desk sold the dip it doesn't expect. In a negotiated block cross, someone sold 30,000 Nov $77.50 puts and collected ≈$13.5M — a bet BHP (≈$83.7) holds above $77.50 (≈7% below spot) through the November expiry, straddling the Aug-18 FY results and the dividend. Selling insurance = you get paid up front, and you're on the hook if it falls hard. Bullish-to-neutral, not a moonshot.

🔵 GEV — GE Vernova: the day's biggest credit — ≈$30M to bet the boom holds. A single-leg auction sold 2,483 June-2027 $800 puts for ≈$30.1M collected, betting GE Vernova (≈$931) stays above $800 (≈14% cushion) into mid-2027. The backdrop is loud: Q2 (7/22) raised guidance, a record 116 GW gas-turbine backlog booked to 2031, Street targets $1,212–1,350. This is a high-conviction income bet on a scarcity story — but a naked short put means the seller eats the downside if the AI-power trade cracks.

⚪ GLW — Corning: ≈$48M that says nothing about direction. Two trades, neither a bet: (1) a $115 call bought via block cross — but a ≈1,000,000-share stock block printed alongside it, so it's delta-hedged long-convexity, not a bullish buy-the-dip; and (2) a deep-in-the-money $310 put that's closing a prior short-put/financing package (≈96% of its price is intrinsic — a financing tell). Corning fell ≈17% on Q2 (7/28) as AI-optical demand outran capacity. The AI-fiber story is real, but this flow isn't playing it directionally.

🐂 INFY — Infosys: a cheap, high-leverage lottery on the AI rebound. Someone paid the offer for ≈83,490 August $15 calls (≈$2.5M) — a fresh, aggressive bullish bet on the Indian IT giant, whose ADR just bounced ≈13% (≈$10.88 → $12.68) after fiscal-Q1 results (7/23) on surging AI/Topaz revenue. But be clear-eyed: the $15 strike is ≈18% out-of-the-money with only ≈3 weeks to run, the Street's targets cluster at $9–$13, and there's no confirmed hard catalyst before the August expiry (next earnings ≈mid-October). Big contract count, small dollars — this is a convex flyer that needs a fast pop to pay off, and most far-OTM near-dated calls expire worthless. High risk, not a core position.

⚪ META — Meta Platforms: a $39M income trade dressed up as a big number. About 25 minutes before Meta reported Q2 after tonight's close, a desk sold 16,000 Jan-2027 $750 calls and collected ≈$39M — and printed a ≈432,000-share stock block right next to it to neutralize the delta. That's a covered-call overwrite: short upside above $750 (≈27% away), long the stock, harvesting rich pre-earnings volatility. It is not a bearish bet on Meta. The print itself proved timely — Meta beat on revenue but missed on EPS as AI capex surged, and the stock slipped ≈4% after hours — but this structure makes money from time and volatility, not from direction. A $39M headline that says "income," not "conviction."

🐻 NBIS — Nebius Group: a 2-day crash hedge on a broken momentum name. Someone paid the offer for ≈33,993 puts expiring this Friday at $120 — ≈20% below the ≈$151 stock — for ≈$4.8M. Important: this is not an earnings play (Nebius reports Aug 12, after these expire), so it's a short-dated crash/tail hedge or a downside-momentum chase on a "neocloud" name already ≈40% off its high, with a ≈28.5% short base and a market in revolt over its $20–25B capex plan. A 2-day, 20%-OTM put needs a violent drop to pay off — high convexity, low odds. Read it as protection or a punt, not a forecast.

👥 How Four Different Traders Might Read Today

🎰 YOLO / momentum trader — Three speculative flyers fit the profile, all high-convexity/low-odds: INFY $15 calls (a cheap bullish AI-rebound lottery, ≈18% OTM, 3 weeks, no hard catalyst), NBIS $120 puts (a 2-day, ≈20%-OTM crash punt on a broken neocloud name — needs a violent drop by Friday), and ETN puts — though note the institution stepped out of the Friday-expiry Eaton leg and kept only the September tail, which is the opposite of a YOLO. All three are mostly premium-burn unless the move comes fast. VFC puts are the one clean bearish print with an actual catalyst behind it, but the ≈17% earnings gap is already banked — chasing after the move is the trap. Everything else (META/GEV/BHP/GLW/BE) is a desk selling or hedging — nothing to chase.

📈 Swing trader — Watch ETN into Friday's earnings, and note what the smart money did: it rolled its downside past the print rather than holding a 2-day binary. You don't have to be the one carrying naked risk over an earnings gap either. VFC is a broken-brand story with a 2027 refi overhang — a bounce is a fade candidate, not a bottom. GEV/BHP short-put sellers are telling you where institutions see support ($800 / $77.50); those floors are levels, not guarantees.

🏦 Premium collector — This was your tape: GEV $800 puts (≈$30M) and BHP $77.50 puts (≈$13.5M) are textbook cushioned income sales into strong fundamental stories. The BE collar shows how to fund protection by selling upside you don't need. Just respect that a naked short put carries full downside — size it, don't max it.

🌱 Beginner — Today's big lesson: a giant premium number is not a "buy" signal. Half of today's dollars were desks selling insurance (they get paid, and take the risk) or hedging (protecting another position). "$30M" or "$48M" tells you size, not direction — you have to read whether it was bought or sold, opened or closed. Start there.

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

⚠️ Risk Control & Patience

Unusual options flow is a starting point for research, not a signal to copy. Today makes the point twice over: the loudest tickets (GEV, GLW, BHP) were desks selling insurance, hedging, or closing — the opposite of "load up." A short put collects premium but carries the full downside; a delta-hedged call has no directional view at all; a "BUY" that closes a position means someone is leaving, not arriving.

Before you act on any of these: (1) confirm open-vs-close with next-day open interest — provisional ⏳ flags can invert (four did on Tuesday's tape; one did on this one — see the ✅ July 29 Resolution block above); (2) size positions so a single earnings gap can't hurt you; (3) never sell a naked put in size you can't afford to be assigned. Patience and position-sizing beat chasing a headline premium every time.

Last updated: July 30, 2026 — the next-day OPRA open-interest snapshot resolved all 12 provisional legs from this session. Eleven confirmed as published (several above target); one inverted (ETN's Jul-31 $350 leg was a roll-close, not an open). The Quick Read's new-directional-money figure was revised ≈$18M → ≈$15.4M, ETN's glance-table row and section were rewritten, and a ✅ July 29 Resolution block was added above. Original publication: July 29, 2026.

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