Ainvest Option Flow Digest - 2026-05-26: $150M Across 7 Names — A $102M AMD Profit-Take, a $24M ELV Short-Call CAPITULATION, Two Big Bears on Korea, and a ZS Put Hours Before Earnings

Last updated: 2026-05-27 — every open/close flag now resolved ✅ via pre-market OI.

A calm read of where large options money went today — and, just as important, what it actually means. Today's tape is dominated by TWO big positions being CLOSED — the $102M AMD is confirmed profit-taking on an existing long (OI dropped −7,527), and the $24M ELV is a confirmed short-call capitulation at a ≈$14M realized loss (OI dropped −1,795). Neither is "smart money loading up bullish." Read these as exits, not entries. Size your reactions to the at-risk capital, not the gross dollar headlines.

The Quick Read

We tracked roughly $150M of notable options premium across 7 tickers (gross — the AMD print is provisional on its open/close status). The mix: a deep-in-the-money call sale that almost certainly closes an existing long position, a deep-in-the-money call buy timed to a regulatory deadline, two distinct bearish bets on Korean equities, an AI-datacenter LEAP scoped to a convertible-cap price, and an earnings-tonight put hedge. The honest version:

• The $102M headline — CONFIRMED as profit-taking, not bearish: AMD — 7,500 contracts of the Aug 21 $380 call sold at $135.90 with the stock at ≈$494. ✅ OI dropped 28,644 → 21,117 overnight (Δ −7,527 ≈ trade size). Definitive STC — the whale closed an existing long-call position to realize ≈$114 of deep-ITM intrinsic value as cash profit.
• A $24M short-call capitulation, NOT a bullish bet: ELV — 2,300 contracts of the Jun 18 $280 call bought at $106.50 to close an existing short position. The smoking gun is in our own archive: the same contract had been STO'd on 2026-03-27 when ELV was near $285. Two months and a 35% rally later (CMS 2027 +2.48% rate decision, Q1 beat, PT raises), the short caller paid $24M to close — a realized loss of ≈$14M on the original 1,600 contracts plus the ≈700 added since. ✅ OI plunged 2,298 → 503 overnight (Δ −1,795), confirming BTC. Signal value INVERTS: short-caller throwing in the towel, not institutional bullish loading.
• A $7.6M synthetic short on Korea's 3X bull ETF: KORU — 107 contracts of a deep-ITM ($1,400 strike vs. ≈$988 spot) Sep 18 put, OI = 0 = brand-new contract. Delta ≈ −0.95 makes this a pseudo-short-stock vehicle with loss capped at the $7.6M premium.
• A $6.6M short put on the AI-memory ETF: DRAM — two STO tranches (15K + 5K contracts) on Aug 21 $50 puts (≈16% OTM), collecting ≈$4.40/contract on the Roundhill Memory ETF. Premium-collection bet that DRAM holds above $50 through three earnings prints (Micron, Samsung prelim, SK Hynix).
• A $4.3M short call on Korean tech: EWY — 1,750 contracts of the Dec 18 $230 call sold on the iShares South Korea ETF (≈16% OTM, Size > OI confirms opening). $230 sits at the upper edge of Goldman/Morgan Stanley KOSPI target bands.
• A $3.9M LEAP positioned to a convertible cap: IREN — 2,000 contracts of the Jan 2028 $110 call bought at $19.39, spot ≈$59.81 (≈84% OTM). The $110 strike matches IREN's convertible capped-call cap price of $110.30 from the May 11 raise — a deliberate "double-from-here" target.
• A $2.3M put bought hours before Zscaler earnings: ZS — 1,800 contracts of the May 29 $187.50 put bought at $12.50. ZS reports Q3 FY26 at 4:30 PM ET tonight. Vol/OI = 733 means brand-new opening flow. Break-even $175 (≈7% drop). 3-day binary.

A large print shows what one participant did with size — not what will happen. Today's AMD headline is most likely a holder cashing chips, not a new bearish bet — that changes the right interpretation completely. Read these as starting points for your own work, size so a total loss of the premium wouldn't hurt, and never chase a print after the move already happened.

Today's Flow at a Glance

Premium figures match each detailed analysis. For DRAM and EWY the dollar amount is credit collected, not capital at risk. For AMD, the gross dollar figure is materially overstated if the trade is STC — most of the $102M is the long position's existing intrinsic value being realized.

The Whale Lineup

1. 🐋 AMD — $102M Whale Looks Like Profit-Taking, Not Bearishness

See why the $380 call sale is most likely closing an existing long position →

• What's happening: 7,500 contracts of the Aug 21 $380 call sold for $135.90 each at 10:32 ET. With AMD at ≈$494 that's $114 deep in the money — about 84% of the option price is pure intrinsic. Vol/OI = 0.26 against an existing 29,000-contract OI = volume folding into the existing book.
• The honest read: This pattern (deep ITM + Vol/OI < 1 + expiration after a known earnings event) is the textbook fingerprint of a holder closing a long position, not opening a fresh naked short. The classifier flagged STO at LOW confidence; we're calling STC ⏳ pending the next-day OI confirmation. AMD is up ≈+109% YTD into Q1 earnings beats, the Meta 6 GW / OpenAI 6 GW / Oracle MI450 deal cluster, and the Aug 4 Q2 print. A whale locking in pre-earnings profit on a position opened months ago at lower prices is the highest-probability story.
• Why it matters: If we frame this as a bearish bet, we mislead readers. If it's actually a profit-taking close, the at-risk capital and signal value are completely different — and the article walks through both interpretations honestly.

2. 💊 ELV — $24M Short-Call CAPITULATION (NOT a Bullish Bet)

See the short-caller throwing in the towel after a 35% rally →

• What's happening: 2,300 contracts of the Jun 18 $280 call bought at $106.50. The original framing of this as a bullish synthetic-stock substitute was WRONG. The cross-reference is unambiguous: the same contract was STO'd on 2026-03-27 when ELV was near $285 — 1,600 contracts for $2.6M premium collected. Today's BUY closes that short.
• The math on the loss: The 3/27 short collected ≈$2.6M. The 5/26 close paid back ≈$10,650/contract. Realized loss on the original 1,600 ≈ −$14.4M. The remaining ≈700 contracts of today's volume represent additional shorts built between March and May that we didn't have in the archive — total close-out value $24M.
• What ran them over: April 6 CMS 2027 MA rate decision (+2.48% vs. +0.09% proposed = ≈$13B sector tailwind), the April 22 Q1 beat ($12.58 vs. $11.03), and UBS/Mizuho raising PTs to $460 and $435. ELV recovered from ≈$285 to ≈$385 in 8 weeks — a ≈35% rally that destroyed the short.
• Signal value INVERTS: A short-caller capitulating at deep ITM is NOT institutional bullish loading. It's often a near-term TOP signal (the late short throws in the towel right as upside exhausts), sometimes a squeeze accelerator. Either way, do NOT read this as "smart money getting long ELV."

3. 🐻 KORU — $7.6M Deep-ITM Put = Synthetic Short on the 3X Korea Bull ETF

See the structured bet against Korean tech without naked short-stock risk →

• What's happening: 107 contracts of a Sep 18 $1,400 put bought at $709.50 with KORU near $988. OI was 0 before this print — fresh open. Delta ≈ −0.95 makes this an options-wrapper synthetic short with the loss bounded at the $7.6M premium paid (vs. uncapped risk if you shorted the 3X ETF outright into a squeeze).
• The catalyst stack: KOSPI hit an all-time high May 15 then crashed -6.12% the same day with $3.7B in foreign net selling. KOSPI forward P/E ≈30× vs. S&P ≈22×. Samsung union back-to-talks Jun 7, SK Hynix Q2 Jul 29, Samsung Q2 late July, Trump-Xi semiconductor summit summer. The Sep 18 expiration covers the entire stack.
• The hidden tailwind: In March 2026, KORU fell -27.83% while EWY (1X Korea) fell only -8.66%. That 3.2× downside ratio (worse than the stated 3× leverage) is volatility drag — daily-reset 3X ETFs systematically destroy NAV in choppy markets, even if the underlying KOSPI ends flat. Free decay tailwind for the put buyer.

4. 🛡️ DRAM — $6.6M Short Put on the AI-Memory Supercycle

See why $50 is "the right" strike on the Roundhill Memory ETF →

• What's happening: Two STO tranches — 15,000 contracts at 11:22 ($4.4M premium) and 5,000 contracts at 11:00 ($2.2M premium), all Aug 21 $50 puts. Spot ≈$60, so the strike is ≈16% OTM. Vol/OI ratios of 625× and 1,875× confirm massive fresh opens. Breakeven ≈$45.60.
• DRAM is the Roundhill Memory ETF (BATS, launched April 2, 2026): SK Hynix ≈24% + Samsung ≈25% + Micron ≈20% combined = ≈73% of the fund. The bull thesis: Q2 2026 DRAM contract prices forecast +58-63% QoQ on top of +90-95% in Q1; HBM sold out through 2026; hyperscaler capex >$600B with 30% to memory.
• The risk to the seller: 49% of the fund's NAV is Korean equities that trade overnight when NYSE is closed. Samsung's preliminary Q2 guidance ≈Jul 7-8 could create a Korean gap-down landing inside expiry. The $50 strike sits exactly on a gamma support wall (total GEX 6.68) — favorable dealer-flow mechanics for the seller, but doesn't backstop an overnight policy or macro shock.

5. 🐻 EWY — $4.3M Short Call as Korea's AI-Chip Cycle Approaches Peak

See the premium-collection bet at the exact Street ceiling →

• What's happening: 1,750 contracts of the Dec 18 $230 call sold at $24.65 ($4.3M premium collected). Spot ≈$198, ≈16% OTM. Size > OI (207) confirms opening. Vol/OI ratio = 8.7.
• Why $230 specifically: Goldman Sachs has KOSPI at 9,000 (≈12% upside), Morgan Stanley at 9,500 base (≈18%) with 10,000 bull (≈24%). EWY at $230 ≈ +16% from spot — almost exactly at the upper edge of the analyst consensus band. The seller is betting the index does NOT break above the Street's high target through year-end while collecting time decay along the way.
• Memory-cycle-peak narrative inside the option's life: SK Hynix already at 72% operating margin (Q1 2026), Samsung Q1 memory profit up nearly 50× YoY, HBM/DRAM consensus peak Q3-Q4 2026 — exactly the window where theta acceleration takes over. Combine that with KRW 36.47T record retail margin debt concentrated in two chips, two BoK rate hikes expected 2H 2026, and live Trump 25% tariff threats, and the seller has a thesis stack.

6. ⚡ IREN — $3.9M Jan-2028 LEAP Targets the Convertible Cap

See the LEAP scoped to the May 11 convertible's capped-call cap price →

• What's happening: 2,000 contracts of the Jan 21, 2028 $110 call bought at $19.39 ($3.9M premium). Spot ≈$59.81. ≈84% OTM, 20-month LEAP.
• The killer detail: IREN's $2.6B convertible notes priced May 11, 2026 included capped calls with a cap price of exactly $110.30 (100% above the May 11 close of $55.15). The whale is positioning for the precise scenario where IREN breaches its convertible-cap protection — a deliberately chosen "double-from-here" target tied to the company's own capital-structure hedging.
• The substance behind the bet: Microsoft $9.7B / 5-year GPU cloud at Childress 750MW with NVIDIA GB300 racks, NVIDIA $3.4B + 30M-share warrant at $70, targeted $3.7B annualized AI Cloud ARR by end-2026 (from $134M run-rate today). Q3 FY26 missed badly ($144.8M vs. $219.87M consensus, -34.1%) as mining is decommissioned for GPU capacity. Analyst split is severe: JPMorgan Underweight $46 vs. Cantor Overweight $61 vs. Street median $80. The LEAP gives the whale 20 months for the execution story to play out.

7. 🛡️ ZS — $2.3M Put Bought Hours Before Earnings Tonight

See the 3-day binary against ZS's +24% pre-print rally →

• What's happening: 1,800 contracts of the May 29 $187.50 put bought at $12.50 at 12:06 PM ET — about 4 hours before Zscaler's Q3 FY26 earnings call (4:30 PM ET tonight). Spot ≈$188 intraday. Vol/OI = 733 with prior OI of just 3 — unambiguously fresh opening flow. Break-even at expiry = $175 (≈7% drop).
• The setup that justifies the bet: Options are pricing ≈±11-13% implied move vs. ZS's 8.25% trailing 8-quarter actual average — options market is paying up for outsized reaction. The stock has rallied ≈+24% in the last 30 days INTO the print, setting a high bar. Recent PT trajectory is downward: Citizens $290 → $210 (May 1), B. Riley $275 → $225 (May 18, paired with upgrade), Morgan Stanley $200 → $155 (Apr 22) flagging "modest Red Canary traction." Tonight is the first full quarter of $4B Red Canary integration disclosure.
• The honest risk for any reader thinking of copying: Post-earnings IV crush. If ZS beats and the stock pops, this put goes to near zero. If ZS sells off but slowly, theta still chews the premium. This is a tight, time-compressed binary — designed for someone who could absorb a 100% loss on the premium and was willing to pay up for the right of a defined-risk downside expression on a name that's been re-rated -47% in seven months.

Read It By Your Style

Not recommendations — just how disciplined approaches might think about today's tape. Risk control first.

• 🎲 YOLO / event trader: ZS is the most extreme event play on the tape — a 3-day binary on tonight's earnings. The catch: implied move ±11-13% means you're paying up for the right of the move. Buying earnings volatility at the top of its implied move band is a classic trap. KORU is a structured directional bet but the premium is huge ($7.6M); not a retail-friendly copy.
• 📈 Swing trader: IREN's LEAP is a directional view but the 20-month horizon means it's really a position-trade, not a swing. ELV is NOT a swing setup the way it first appeared — a short-call capitulation can mark near-term tops or fuel short-squeeze continuations, and either way it isn't a clean fresh-conviction signal. If you want exposure to ELV recovery, the underlying ELV stock pages and Q2 print in late July are cleaner reference points than this exit print.
• 💵 Premium collector: DRAM and EWY are today's two pure premium-collection structures. The lesson: both sellers picked strikes anchored to specific reference points — DRAM $50 sits on a gamma support wall, EWY $230 sits at the analyst-consensus ceiling. That's the discipline — not just "sell volatility," but sell at structurally meaningful levels. Recognize the tail risks (DRAM = Korean overnight gap; EWY = supercycle running another 12-18 months).
• 🌱 Just getting started: Today's lesson is read the open/close, not just the gross dollars. The two biggest prints — AMD ($102M) and ELV ($24M) — are BOTH likely CLOSING existing positions, not opening new ones. ELV is confirmed: a short-caller paid $24M to close a position opened 8 weeks ago for $2.6M (a ≈$14M realized loss). The "$102M AMD whale" is most likely a holder cashing existing long-call chips. DRAM/EWY's $10.9M combined "premium" is actually $10.9M received, not paid — credit collected, not capital at risk. ZS's $2.3M is a 3-day earnings binary that could be worth ≈$0 by Friday. Read the direction and at-risk capital, not the headline number.

Upcoming Catalysts — Event vs. the Option That Plays It

Memorial Day yesterday (May 25) was closed; today's tape is the first full session of the week. Watch for the AMD provisional ⏳ flag resolution in tomorrow's pre-market OI update — that single data point may flip the AMD narrative.

Risk & Reminder

Large prints look like signals; they're often just one institution rebalancing a book. The four most important things to remember as you read these:

1. Open/close matters more than direction. AMD shows it bluntly — a "$102M call sale" reads completely differently if it's STC (closing a profitable long) vs. STO (opening a fresh naked short).
2. Gross dollars are not at-risk capital. Spreads, credit-collection STOs, and deep-ITM long calls all change the right way to size against the print.
3. Time decay is brutal on short-dated trades. ZS's earnings put will lose value rapidly even if the thesis is correct — IV crush is real.
4. Strike geometry hides the intent. Deep-ITM calls/puts can be synthetic-stock substitutes, not directional bets. Deep-OTM LEAPs anchored to a known reference price (like IREN's convertible cap) tell you more about the trader's framework than a generic "bullish bet" headline.

Read every detailed analysis, never copy without doing your own homework, size so a total loss of the premium wouldn't hurt, and remember that the best traders are the most patient — not the most reactive.

Complete Link Directory

• AMD — $102M whale, profit-taking probable
• [ELV — $24M short-call CAPITULATION (≈$14M realized loss closing the 3/27 STO)](https://optionpilot.ainvest.com/idea/flow/ELV?date=2026-05-26)
• KORU — $7.6M synthetic short on Korea 3X bull ETF
• DRAM — $6.6M short put on the AI-memory supercycle
• EWY — $4.3M short call at the Street's Korea-target ceiling
• IREN — $3.9M Jan-2028 LEAP at the convertible-cap price
• ZS — $2.3M put hours before earnings tonight

Disclaimer: This newsletter analyzes large unusual options activity reported by exchange tape. Nothing here is investment advice. Options trading carries significant risk; you can lose 100% of premium paid (or more for short positions). Always do your own research and size positions to your risk tolerance.

Last updated: 2026-05-26