Ainvest Option Flow Digest - 2026-07-10: 🧩 The ≈$94M of "Flow" That's Really Financing — and the One $10M Trade That's Actually a Bet

❗ UPDATED July 13, 2026 — two reads in this digest INVERTED on the next-day OI check. The OPRA open-interest snapshot for this session is now in. 💵 MOS's "$13M put sale" was an EXIT, not a bullish carry trade (OI FELL 12,582 → 10,073 — an opening sale cannot reduce open interest, so the seller was closing a long put). 🐺 WOLF's "bull call spread" was never a spread (the $60 leg's OI COLLAPSED 25,696 → 12,101 — those calls were being closed, not shorted; it's a roll DOWN from $60 to $45). Both articles have been corrected. The other eight names held. Details below and in each ticker page.

🎯 Quick Read: ≈$117M Tracked — But ≈$94M Is Financing; Only ≈$23M Is a Real Directional Bet, and META Owns the Clean Signal

Today's tape is a lesson in reading structure over size. We tracked ≈$117M across 10 names — but ≈$94M of it isn't a directional bet at all. The two biggest tickets — AMD ≈$34M (a deep-ITM call cross plus a 0DTE financing box) and MSTR ≈$42M (a deep-ITM call roll at parity) — are pure financing/synthetic structures, joined by MOS $13M, GRAB $2.6M and DOCN $2.0M. Both are deep-in-the-money financing/premium structures: AMD is a negotiated block cross against a 17,500-lot existing line, and MOS is a deep-ITM LEAP put sale (a credit collected, synthetic-long carry). Strip the financing out and the genuine directional flow is just ≈$23.1M — split between bullish (🌋 META $10M, 💳 COF $4.1M, 🐐 WOLF $2.7M, 🥇 AEM $1.8M) and defensive (🛡️ GS $4.5M put roll into earnings).

• The one clean bet: 🌋 META $10M — an aggressive lit BUY of Jul-31 $685 calls that lifted the ask (not a cross, not an auction — real liquidity-taking). Fresh open (size 2,980 vs OI 168), 2 sessions before July 29 earnings. This is the day's only trade where the tape proves the direction: bullish.
• The biggest headline, the smallest signal: 🤝 AMD ≈$34M across TWO financing trades — a $28M deep-ITM $450 call block cross (SELL, against an existing 17,500-lot line) AND a ≈$6.2M 0DTE $515/$530 call spread that paid ≈full strike width (a delta-hedged financing box). Neither is a directional AMD bet.
• The parity roll: 🟠 MSTR ≈$42M — a deep-ITM call roll at parity on Strategy/MicroStrategy: close the expiring 0DTE $35 calls, open Jul-17 $30 calls, all at ≈pure intrinsic (delta ≈1, ≈no time value). That's a synthetic-long / financing roll tied to MSTR's liability management (it just made its first bitcoin sales since 2022 to cover preferred dividends) — not a directional bitcoin bet.
• The $13M EXIT (❗ INVERTED July 13): 💵 MOS $13M — a deep-ITM Jan-2028 $30 put sale. We published this as a bullish premium-collection carry ("getting paid to own Mosaic near $19"). Next-day OI FELL 12,582 → 10,073 (−2,509) — proving the seller was closing a long put line they already owned (STC), not writing a new one. The $13M was liquidation proceeds, not premium income. No new short put exists; there is no bullish carry trade here.
• The synthetic long: 💵 GRAB $2.6M — a sale of 30,000 in-the-money Jan-2027 $4.50 puts (a ≈$2.6M credit collected). Selling ITM puts is a synthetic-long / income stance — the seller effectively agrees to own Grab near $3.62, below spot. A block cross, so direction is a lean, not proof.
• The financing package: 🤝 DOCN $2.0M — a slightly-ITM $128 call printed as a floor stock+option combo (an equity leg trades alongside the option → likely delta-hedged/buy-write). Not a directional call buy; the Jul-17 expiry lands before the Aug 12 earnings, reinforcing a financing/hedge read.
• The hedge: 🛡️ GS $4.5M — a protective put roll, up in strike and out in time, blanketing the July 14 (before-open) earnings print. Goldman fell on its record Q1 — this is insurance near all-time highs, not a naked bearish swing.
• The roll-down (❗ INVERTED July 13): 🐺 WOLF $2.7M — published as a Jan-2027 $45/$60 bull call spread. Next-day OI on the $60 leg COLLAPSED 25,696 → 12,101 (−13,595) — a fresh short cannot reduce open interest, so those $60 calls were being CLOSED. Paired with a confirmed fresh open at $45 (+21,920), this is a roll DOWN: a holder abandoned the $60 strike and re-struck at $45. Still net long calls (still bullish), but it is a retreat to a reachable strike, not a capped spread — and upside is NOT capped at $60.
• The bank bet: 💳 COF $4.1M — a bullish-lean buy of Sep-18 $210 calls, a leg of a worked multi-leg auction (fresh open by size). Spans Capital One's July 21 earnings and the July 27 Discover-integration migration start — but as an auction leg, the net direction is a lean, not proof.
• The gold LEAP: 🥇 AEM $1.8M — a bullish-lean buy of far-OTM Jun-2027 $175 calls (≈19% OTM) on Agnico Eagle, a block cross. A leveraged multi-quarter bet on a gold recovery (gold round-tripped from ≈$5,500 to ≈$4,140 in 2026) — but a cross can't confirm the side.

Headline figure is NET/structural (long premiums paid − short premiums collected); gross overstates the spreads. Post-OI revision (July 13): MOS's $13M is now classed as a position exit rather than a financing carry, and WOLF's $2.7M as a roll-down rather than a fresh spread — the dollar totals are unchanged, but the reads are not. Genuine directional flow ≈$23.1M (META + COF + AEM bull-lean, WOLF bull-lean via roll-down, GS hedge); AMD + MSTR + GRAB + DOCN are financing/non-directional (deep-ITM crosses, parity rolls, put sales — GRAB leans bullish, unprovable).

---

🔁 OI Review — Last Session's Provisional Flags, Now Resolved (covers 2026-07-09)

The next-day OPRA open-interest snapshot for the 2026-07-09 session is in (posted 2026-07-10 pre-market). We pulled it for all 9 legs across 7 names and resolved every ⏳ provisional flag. The headline: six names opened as read — but 🧬 RARE's "$3M put diagonal" was never a diagonal at all, and 🤝 NVDA's "21x prior OI" whale evaporated once we found the real open-interest baseline. A big SELL headline is never a short until next-day OI proves it opened.

🔄 Inversions

• RARE — "bearish put diagonal (fresh short leg)" → BEARISH PUT ROLL, up & out. We flagged the Sep 18 $22.5 "short" put as provisional because size (6,000) exactly equalled prior OI (6,000). Next-day OI FELL 6,000 → 1,227 (−4,773). Open interest only declines when both sides of a print are closing, so the seller must have been long and liquidating — a Sell-To-Close, not a fresh short. Paired with the Oct 16 $30 put's flawless open (OI 1 → 6,001, +6,000), this was one packaged order doing a roll: close 6,000 September puts, open 6,000 October puts. Corrected thesis: there is no short leg, no premium harvested, and no assignment risk — the trader paid a $2.98M net debit to roll up in strike and out in time (now spanning both the Aug 23 DTX401 and Sep 19 UX111 PDUFAs). Capital genuinely at risk is the full $4.95M of long puts, not a $3.0M spread. Conviction went up after an adverse move.

✏️ Corrections

• NVDA — prior open interest was 35,052, not 100. Our scanner reported an OI baseline of 100 on the Jan-2028 $250 LEAP call; the true figure was 35,052 — one of the most crowded LEAP strikes on the board. That bad number turned a routine 2,000-lot block into a phantom "21x prior OI" whale; the real ratio is ≈0.06x. Because size sat far below prior OI, open vs close was never provable from the tape. Next-day OI rose 35,052 → 36,460 (+1,408): a partial net open (≈70% new contracts). Unusual Score revised MODERATE-HIGH → MODERATE. Direction still unprovable (a cross).

✅ Confirmations

• MRVL — both short-call legs OPENED ($250: 4,103 → 6,197, +2,094; $230: 1,245 → 3,221, +1,976). The $18.2M overwrite was a genuine fresh short on both layers.
• MU $825 put cross OPENED (125 → 5,546, ≈98.6% absorbed). HUM $360 put cross OPENED (13 → 1,012, near-perfect).
• EEM $69 call cross OPENED (3,073 → 18,179, +15,106) — a companion 3,000-lot cross printed at the identical price. Direction still unprovable.
• SPY parity floor combo OPENED exactly (2 → 2,002, +2,000) — and it changes nothing. A perfect open on a deep-ITM-at-parity print is still not a bullish bet.

What changed: RARE went from a "$3.0M net-debit diagonal with a premium-collecting short leg" to a $4.95M long-put roll with no short leg at all — more bearish, more capital at risk. NVDA's "21x OI whale" became a normal block on a crowded strike, only ≈70% new risk. The standing lesson cuts both ways: when trade size is at or below prior open interest, no order type is provable from the tape — and when the OI number itself is wrong, every ratio built on it is wrong too.

---

📊 The Glance Table

---

🔬 The Ten Positions — Decoded

1. 🤝 AMD — The $28M Deep-ITM Call Block Cross (financing, not a bet)

SEE BOTH AMD FINANCING TRADES →

• Flow: ≈$28.0M — sold 2,000 Oct-16 $450 calls (deep ITM, ≈$99 intrinsic) at $140.30 as a block cross near the bid. Spot $549.05. Prior OI 17,494, size 2,000.
• Read: a cross means a broker matched a known buyer and seller off the open book — there is no aggressor and no proven direction. With size far below the existing 17,500-lot line, this reads as position management / a call overwrite / financing, not fresh conviction. The biggest dollar figure of the day carries the least directional signal. Q2 earnings Aug 4; Advancing AI July 22–23 — but this print isn't positioning for them.

2. 🟠 MSTR — The ≈$42M Deep-ITM Call Roll at Parity (financing, not a BTC bet)

SEE THE MSTR PARITY ROLL →

• Flow: three legs at 15:59:11 as a worked multi-leg auction, all deep in-the-money at parity (option ≈ pure intrinsic): sold 4,609 0DTE $35 calls (≈$27M, closing the expiring line, size = OI) and opened Jul-17 $30 calls (sold ≈3,410 / bought ≈1,012, prior OI just 35). Spot $94.64.
• Read: deep-ITM-at-parity calls have delta ≈1 and ≈no time value — they behave like the stock itself, so rolling one from today to next week is a financing / synthetic-long roll, not a directional bet. It fits MSTR's moment precisely: the company just made its first bitcoin sales since 2022 (≈$216M on July 6) to cover preferred dividends, with the stock trading ≈at or below its bitcoin NAV. Q2 earnings July 30. There is no bullish or bearish signal to copy here — it's balance-sheet plumbing.

3. 💵 MOS — The $13M Deep-ITM LEAP Put Sale ❗ INVERTED: it was an EXIT

SEE THE $13M MOSAIC PUT SALE →

• Flow: ≈$13.0M CREDIT collected — sold 12,253 Jan-2028 $30 puts (deep ITM: spot $21.22, ≈$8.78 intrinsic) at $10.90 via a price-improvement auction. Prior OI 12,582 (flat for days), size 12,253.
• ❗ CORRECTED READ (July 13): we published this as a bullish premium-collection carry — "getting paid to own Mosaic near $19." The next-day OI snapshot inverted it. Open interest at the strike FELL from 12,582 to 10,073 (−2,509), and this print was 12,253 of the strike's 12,255 total daily volume, so there is nowhere else to attribute the move. An opening sale can never reduce open interest. The seller was selling to CLOSE (STC) a long put line they already owned. The $13M was liquidation proceeds, not premium income — nobody wrote a new put and nobody agreed to buy MOS at $19. ≈9,744 contracts simply transferred to new long holders. Directionally this is close to a non-event: if those puts hedged long MOS stock, dropping the hedge is mildly bullish; if they were an outright bearish bet, closing is a neutral unwind. The tape cannot distinguish, and we won't pretend it can. Q2 earnings Aug 3.

4. 🌋 META — The $10M Aggressive Lit Call Buy (the real bullish bet)

SEE THE $10M META EARNINGS CALL →

• Flow: ≈$10.0M — bought 2,980 Jul-31 $685 calls (≈1.7% OTM) at $35.00, lifting the ask in the lit market. Spot $673.40. Prior OI 168, size 2,980 → Vol/OI ≈17.9, a proven fresh open.
• Read: this is the day's cleanest signal — a trade where the tape proves direction (real liquidity-taking, not a cross). It's a levered earnings bet: Q2 prints July 29 after close, the calls expire July 31 (two sessions later), and META has run ≈+24% into the event. The whole $35 is extrinsic — all-or-nothing on the post-earnings move. Counterweight: the same ≈$130B capex overhang sank the stock ≈8% on last quarter's beat, so sell-the-news risk is live.

5. 🛡️ GS — The $4.5M Protective Put Roll (insurance into July 14 earnings)

SEE THE $4.5M GOLDMAN PUT ROLL →

• Flow: ≈$4.5M net debit — bought 1,318 Aug-21 $1,050 puts (fresh open; real prior OI just 44) and sold 1,026 Jul-17 $1,030 puts as a multi-leg auction. Spot $1,054.78.
• Read: a put roll up-and-out — closing/shedding near-dated protection and opening higher-strike, later-dated protection. Goldman reports July 14 before the open; the sold Jul-17 put expires just after the print while the bought Aug put carries protection ≈5 weeks past it. With GS near its all-time high and having fallen on a record Q1, this is well-timed downside insurance, not a naked bearish view.

6. 💳 COF — The $4.1M Bullish-Lean Call (a worked auction into July 21 earnings)

SEE THE $4.1M CAPITAL ONE CALL →

• Flow: ≈$4.1M — bought 4,800 Sep-18 $210 calls (≈4.3% OTM) at $8.52 as a leg of a worked multi-leg auction. Spot $201.33. Prior OI 2,427, size 4,800 → Vol/OI ≈2, a fresh open by size.
• Read: the visible leg is an OTM call bought with a buy-lean (73% across), but on an auction the per-leg aggressor is unreliable and the paired leg isn't cleanly visible — so this is a moderate bullish lean, not proof. The setup is supportive: Capital One reports July 21, the Discover-integration migration starts July 27 (≈$2.5B synergy thesis), credit is improving (card charge-offs 5.1%→4.82%), and consensus target ≈$259 sits ≈29% above spot. The Sep-18 tenor spans both catalysts.

7. 🐺 WOLF — The $2.7M "Bull Call Spread" ❗ INVERTED: it was a ROLL-DOWN

SEE THE $2.7M WOLFSPEED SPREAD →

• Flow: ≈$2.7M net debit — bought 12,500 Jan-2027 $45 calls (fresh open) and sold 12,500 Jan-2027 $60 calls as a multi-leg auction. Spot $34.61.
• ❗ CORRECTED READ (July 13): we published this as a defined-risk bull call spread targeting $60. The next-day OI snapshot says it was never a spread. The $45 leg opened cleanly (4,121 → 26,041, +21,920) — but open interest on the $60 leg COLLAPSED (25,696 → 12,101, −13,595). A sale that opens a new short call creates open interest; it cannot destroy it. Those $60 calls were being CLOSED. The economics are a roll DOWN: a holder who was long ≈25,700 Jan-2027 $60 calls sold out and re-struck at $45, paying ≈$2.7M to move down the ladder. What changes: there is no short $60 leg, so upside is not capped at $60 and the "≈6x max payout" math is void. Still net long calls (still bullish) — but this is a retreat to a reachable strike, not conviction that WOLF reaches $60. They abandoned that target. Q4 earnings Aug 19; ≈$2.5B CHIPS/Apollo stack still in play. High-IV, speculative. (Proven: both OI moves. Inferred: that the same account sits on both legs — OPRA publishes no account IDs.)

---

8. 💵 GRAB — The $2.6M ITM LEAP Put Sale (getting paid to be long)

SEE THE $2.6M GRAB PUT SALE →

• Flow: ≈$2.6M CREDIT collected — sold 30,000 Jan-2027 $4.50 puts (in-the-money: spot $3.93) at $0.88 as a block cross near the bid. Prior OI 3,289, size 30,000 → Vol/OI ≈9, a proven fresh open.
• Read: selling in-the-money puts is a synthetic-long / income-financing stance — the seller pockets ≈$2.6M and effectively agrees to own Grab near $4.50 − $0.88 ≈ $3.62 (below today's price). It lines up with a fundamentally improving name: Grab turned profitable in Q1, is executing a $400M buyback, and carries free GoTo-merger optionality (consensus ≈$6.17 vs $3.93). Structurally bullish-lean — but a cross can't confirm the customer side. Q2 earnings ≈Aug 20.

9. 🤝 DOCN — The $2.0M Floor Stock+Option Combo (financing, not a call buy)

SEE THE $2M DIGITALOCEAN PACKAGE →

• Flow: ≈$2.0M — a slightly-ITM Jul-17 $128 call (spot $131.67) printed as a floor stock+option combo at $9.40, with an equity leg trading alongside the option (prior OI 0).
• Read: a stock+option package worked on the floor reads as delta-hedged / buy-write / conversion financing, not a directional call buy — and %-across can't signal aggression on a floor print, so direction is unprovable. The tell: the Jul-17 expiry lands ≈4 weeks before the Aug 12 earnings (DigitalOcean already pre-announced a record Q2 on July 7), so it doesn't even span the catalyst. A financing/hedge package, not a bet on the print.

10. 🥇 AEM — The $1.8M Gold-Recovery LEAP Call (bullish-lean, far-OTM)

SEE THE $1.8M AGNICO EAGLE CALL →

• Flow: ≈$1.8M — bought 1,000 Jun-2027 $175 calls (≈19% OTM, ≈11-month LEAP) at $17.75 as a block cross. Spot $146.78. Prior OI 305, size 1,000 → Vol/OI ≈3.3, a fresh open by size.
• Read: a far-OTM, long-dated call on a senior gold miner is a leveraged multi-quarter bet on a gold recovery. Gold round-tripped in 2026 — a record ≈$5,500/oz in January, then ≈−26% to ≈$4,140 by July, dragging AEM from the $250s to ≈$146 (sell-side year-end gold targets still sit far above spot). The $175 strike is well below AEM's 52-week high and its ≈$234 consensus target. Bullish-lean — but a block cross can't confirm the customer side. Q2 earnings July 29.

📈 One-Year Performance — All Ten Names

Tenor tags: 🗓️ Near-term: META (Jul 31), GS (Jul 17 / Aug 21), DOCN (Jul 17), MSTR (Jul 10 / 17), AMD 0DTE (Jul 10) · 🍂 Quarterly: AMD (Oct 16), COF (Sep 18) · 🚀 LEAP: WOLF (Jan 2027), GRAB (Jan 2027), AEM (Jun 2027), MOS (Jan 2028)

---

🧭 The Takeaway

The dollar leaderboard and the conviction leaderboard point in opposite directions today — and the next-day OI check drove that home harder than we expected.

Two of our reads inverted, and both inverted the same way: a big SELL that looked like an OPEN turned out to be a CLOSE. MOS's "$13M premium collected" was a position exit (OI −2,509). WOLF's "short $60 leg" was a liquidation of long $60 calls (OI −13,595), turning a bull call spread into a roll-down. Neither could have been called from Friday's tape alone — in both cases the trade size sat below the existing open interest, which is precisely the condition under which open vs. close is unprovable. That's why we flag ⏳ and come back.

What survived intact: META's $10M lit call buy is still the one trade the tape proves is a bet (bullish, fresh open 168 → 3,308, into July 29 earnings). AMD's $28M cross resolved as a clean CLOSE (OI −2,000, exactly the block size) — vindicating the call that the day's biggest headline carried the least signal. GS's put roll and MSTR's parity roll both confirmed exactly as read.

The standing lesson, now with two fresh scars: a big BUY headline isn't conviction and a big SELL headline isn't bearishness. "SELL" can mean writing a new option (opening) or dumping one you own (closing) — opposite meanings, identical print. Only the next morning's open interest tells you which. ✅ This session's OI check is complete; the resolutions above are final.

Not investment advice. Options carry risk; do your own research.

---

Last updated: July 13, 2026 — next-day OPRA open-interest resolution applied to the 2026-07-10 session. 2 inversions (MOS, WOLF); 8 names held.