Huawei's CloudMatrix 384 does not mean China beat NVIDIA on efficiency. It is a structure that bypasses the GB200 NVL72 in aggregate by lashing together 384 Ascend chips, drawing 559kW of power, and connecting them with massive optical links. The real alpha is not in direct Huawei suppliers but in the inspection, metrology, and high-difficulty consumables demand created when China is forced to scale up its own fab, equipment, and consumables investment. 티씨케이, 넥스틴, and 코미코 are not the same China theme. 티씨케이 is a confirmed consumables recovery play with 1Q26 China normalization in the numbers. 넥스틴 is a bet on 2H26 China PO resumption. 코미코 is global foundry quality with a China option attached.
The May 14-15 Trump-Xi summit in Beijing produced a managed truce, not a grand bargain. The more important signal came outside the room: the U.S. cleared H200 sales to China, but no chips have shipped because China has not allowed purchases to proceed. At the same time, Huawei's CloudMatrix 384 shows how China can offset weaker chips with more chips, more power, and more optical interconnects. ITIF estimates that in a full semiconductor decoupling scenario, U.S. firms could lose USD 77B of China-related sales in the first year, while South Korean firms could gain USD 21B. For Korea, the short term is favorable for memory and AI components; the long term brings standard fragmentation and dual-supply-chain costs.
Pearl Abyss's 1Q26 results proved Crimson Desert's commercial success and margin power. Meritz estimates 3.089M recognized Crimson Desert copies in 1Q, implying an accounting ASP of about KRW 86,300 per copy on KRW 266.5B in Crimson Desert revenue. Applying the same ASP to company 2Q Crimson Desert guidance of KRW 224.2B-276.5B implies 2.6M-3.2M recognized copies, with a midpoint near 2.9M. A 3M-recognized-copy 2Q is therefore a reasonable base case. But the market debate has already moved from a 2Q cliff to a 2027 earnings cliff. The next re-rating depends on whether DLC, Asia/China expansion, platform extension, capital return, or DokeV visibility can lift 2027 operating profit from the KRW 200B range toward KRW 300B+.
Samsung Electro-Mechanics is neither a stable MLCC quality compounder like Murata, nor yet a proven FCBGA scarcity leader like Ibiden. But it is a high-beta challenger that simultaneously holds both axes. The current share price of ₩1,010,000 already prices in a 2027E PER of 40.5x — meaning this is not a stock you buy because it is cheap, but one that won't look expensive until the numbers move higher. 1Q26 OPM 8.7%, Package revenue YoY +45%, 2027E Package OPM estimated at 17.9%. Murata FY3/27 OPM 19.4%, Ibiden FY2027 OPM 23.1%. The gap remains wide, and that gap is the room for additional upside. ₩1,200,000–₩1,400,000 opens when 2027E OP is revised up to ₩2.7 trillion or more; ₩1,500,000 opens upon OPM approaching 20% + 2028 visibility + Package LTA confirmation.
Markets wobbled in the latter half of last week. This was not a simple rate shock — it was a five-layer complex risk-off in which an oil price surge, rising long-term real rates, dollar strength, crowded AI positioning, and weak Chinese credit all hit simultaneously. Recovery is possible, but it begins not with an immediate Fed cut, but in this sequence: Hormuz de-escalation → oil decline → long-rate stabilization → dollar weakness → restoration of foreign inflows into Korea. The key is the 7 recovery triggers and the stealth-easing toolkit that can lower market rates without a Fed cut.
Google I/O begins with the keynote at 2:00 a.m. KST on May 20, followed by NVIDIA's FY2027 Q1 results at about 5:20 a.m. KST and conference call at 6:00 a.m. KST on May 21. Google I/O is about AI use-case expansion; NVIDIA earnings are about AI infrastructure CAPEX confirmation. For Korean semiconductors, NVIDIA's Q2 guide and 75% gross-margin durability matter more directly.
Both stocks exploded in 1Q26. Hana Micron posted operating profit of KRW 72.0bn (+29.7% above consensus), OPM 14.2%. Jeju Semiconductor posted KRW 67.1bn (+1,714% YoY), OPM 37.2%. The day after results, Hana Micron rose +18.6% and Jeju Semiconductor +8.9%. On the surface, both look like 'AI-era memory back-end beneficiaries.' But the nature of the two surprises is different. Hana Micron reflects structural improvement — a repriced cost-pass-through mechanism at Vina and a margin step-up at the Brazil subsidiary. Jeju Semiconductor reflects a cyclical windfall — LPDDR4X supply scarcity compounded by tariff-driven front-loading. Same word, 'surprise.' Very different durability.
Hyundai Mobis at KRW 629,000 already prices in most of the Atlas 30K-units/year actuator-supply scenario. What is NOT yet in the price: gripper / sensor / controller expansion, external (non-HMG) customer wins, and a re-rating of the indirect 10.9% Boston Dynamics stake. On 30K Atlas units, Mobis's robot-parts revenue band is KRW 2.1–3.0T, operating profit KRW 170–360B, and the segment EV is KRW 3.4–7.2T. Today's quote already embeds roughly KRW 5T of robot optionality. The stock is therefore not 'cheap' — it is in a 'verification window.' New buyers: Wait. Holders: Hold. Additions should follow confirmation of scope, units, and margin. The biggest single trigger is non-HMG customer announcements (a USD ~190K humanoid market scenario) and gripper / sensor expansion disclosures.
AI-RAN could come up on next week's NVIDIA earnings call, and the market's reflex will be to buy SK Telecom. But the economic value created by AI-RAN flows to GPUs, HBM, vRAN equipment, RU/RF components, and optical interconnects before it reaches the operator. Here is why a Korean AI-RAN supply-chain basket is a more rational approach than a single-name SKT position.
As of May 16, 23:20 KST, Crimson Desert CCU is 66,448, Steam global sales rank #18, US #19, Korea #7, China #16 — all recovered. But average CCU vs the same window last weekend is still -12%. The stronger signal this weekend is not 'more users' — it is 'higher sales-rank and review quality.' New-review positive rate is 94.0% vs last weekend's 82.1%, suggesting Patch 1.07 has eased most of the post-1.06 negative pressure. Cumulative reviews at 151,847 imply a central units estimate of ~5.81M, with a 6M ETA around May 27–29. The most interesting new signal — Steam China sales rank is structurally stronger than global over the past week, with a 7-day average of China #17.3 vs global #23.4 (~+6.2 rungs better, China leading on 98.6% of observations). This is not 1–2 day noise — it is a week-long pattern, a 'China optionality is waking up' signal. Read: Patch 1.07 defended the long-tail thesis; the China signal adds another layer. Next step set — formal 6M, DLC concretization, capital-return, and a real China-mainland signal all need to land together.