SKC owns 45% of a separately-listed AI test-socket maker — that stake alone is 40% of the market cap.

• The mark. SKC's 45.03% stake in KOSDAQ-listed ISC marks at ₩2.42T against SKC's ₩6.05T market cap. ISC printed ₩60B operating profit on ₩220B revenue at 27% margin in FY25, then +236% YoY operating profit at 35% margin in 1Q26 on AI/HBM tester demand.
• The multiple. ISC trades at 95× trailing P/E and ~22× the implied EV/Sales of LEENO Industrial, its closest pure-play peer. The premium pays for AI/HBM design-in lock SKC management does not control — if AI capex disappoints or HBM share rotates to Micron, the largest single mark in the SOTP would compress with nothing else changing.
• The hidden bet. The 9-analyst consensus target of ₩105,864 only closes mathematically if ISC compresses 25–30% from its KOSDAQ mark — but no covering analyst has explicitly underwritten that view. A short on SKC is implicitly a short on the AI cycle through ISC.

First positive consolidated EBITDA in eleven quarters. The bear case is whether 5 August holds it.

Three engines moved together for the first time in two years — chemicals swung positive on a Mid-East PG-spread spike, the Malaysia copper-foil plant turned standalone EBITDA-positive on volume pull-through, and ISC posted record AI-socket margins. The bear's wedge is that 4Q25 booked an unquantified ₩316.6B write-down of "manufacturing-equipment optimization" right before the rights raise, flattering the QoQ comparison. The 5 August 2Q26 print is the first clean comparison — a second consecutive positive print with chemicals OP retained validates the inflection; reversion to negative validates the artifact reading.

₩1T rights offering lists 8 June; new CEO has one earnings cycle of credibility.

• Mechanical 22% dilution lands 8 June. SK Inc. (40.64% controller) committed to 120% of pro-rata; ESOP demand survey 132% oversubscribed. ₩590B routes to Absolics + AMAT collaboration; ₩410B retires debt — D/E drops from 233% to 142% post-deal.
• The CEO conflict cleared. Kim Jong-woo replaced Park Won-cheol on 26 March 2026, ending the prior CEO's dual role at SKC parent and Absolics — a structural conflict now resolved. New mandate: "stability, recovery, growth." CFO Park Dong-ju was hired from SK Inc. portfolio planning in late 2025.
• Four years of credibility damage. Six of ten valuation-relevant promises since 2022 were missed or quietly abandoned — the 250kt copper-foil capacity target, the silicon-anode roadmap, and the ₩2.5T net-debt target. The asset-disposal program (₩893B in 2025 across CMP Pad, FCCL, blank mask) did deliver. The new CEO inherits both records.

+75% in six weeks at decade-high vol, into a 22% mechanical dilution and an UNDERPERFORM consensus.

• The blow-off. 30-day realized volatility hit 117% on 7 May — the highest in ten years, 33–48 points above any prior peak (2021-Q4 79%; 2024-Q2 80%; 2025-Q1 88%). Price sits 49.5% above the 200-day SMA, RSI ~70. May 7 traded 2.76M shares — 6.6× the 50-day average.
• The institutions disagree. 9 sell-side analysts cover with mean UNDERPERFORM and 12-month target ₩105,864 (-34% from spot). Nomura at "Reduce" / ₩90,000. JPMorgan at Sell / ₩72,000 (cut from ₩84,000 on 2 March). Goldman Sachs at Sell / ₩81,000 (28 April). Macquarie maintained Hold / ₩130,000 on 29 April — no covering analyst has upgraded into the rally.
• The arithmetic trap. ₩105,864 × 37.9M shares = ₩4.01T market cap. Subtract the ₩2.42T ISC mark and "everything else" residual = ₩1.59T — which only closes if consensus is also implicitly short the AI cycle through ISC. Either the targets quietly assume ISC compresses, or they are mathematically inconsistent with the listed mark.