· 个股论点

伊朗袭击致LNG暴涨引发韩股暴跌,但AI需求强劲使三星/SK海力士可转嫁成本,视为买入机会。

涉及标的:

中文翻译

关于 $EWY / KOSPI 暴跌以及这是否是三星/SK海力士的买入机会的思考: 这看起来是一个明确的买入机会。 尤其是因为投资者误解了2026年并非2022年(俄乌冲突)的局面。AI从根本上改变了内存需求,与2022年消费电子低迷时期不同,当时三星/SK海力士不得不自行承担运营支出(OPEX)。 韩国历史上最大单日跌幅的主要催化剂是伊朗无人机袭击卡塔尔能源公司(Ras Laffan)综合体。 卡塔尔约占全球液化天然气(LNG)供应的~20%。 结果: -> 荷兰TTF(欧洲天然气):+46% -> 亚洲现货LNG:+39% -> 美国天然气:+4.16%(正常) LNG飙升是下跌的催化剂,因为韩国严重依赖进口LNG,暴露度极高。 出于对通胀性能源冲击的恐惧,投资者激进抛售韩国科技股,叠加10倍杠杆三星/SK海力士的去杠杆效应。 导致: -> 三星电子下跌9.79%。 -> SK海力士股价暴跌11.12%。 -> $EWY(韩国期货)今日下跌14.49%。 人们担心重演2022年俄乌战争导致LNG价格上涨,韩国电力公社(KEPCO)停止吸收亏损后,韩国工业电价上涨约70%。 仅2022年,三星国内设施就消耗了约28,000 GWh的电力。 电价上涨使三星年度运营支出(OPEX)增加约2万亿韩元(超过14亿美元),SK海力士增加超过1万亿韩元。 这种前所未有的运营支出激增发生在最糟糕的时机。 2022年底,半导体行业进入残酷的周期性低迷。后疫情时代PC和智能手机需求暴跌,导致DRAM和NAND闪存严重供过于求。 然而,2026年完全不同: 与2022年低迷期面临能源成本上升相比: AI对内存的需求完全史无前例。27日有报道称三星/SK海力士再次上调DRAM价格。 $SNDK的NAND需求如此之高,以至于产能分配需要提前3年预订。 需求看不到尽头,直到2028年都没有缓解迹象。 相比之下,2022年笔记本电脑和智能手机的消费端已经面临低迷,内存制造商无法转嫁成本(并承担了巨额OPEX账单)。这导致了上次的大幅修正。 由于2026年全球AI芯片短缺如此严重,他们将无缝地将高昂的公用事业账单转嫁给供应链下游。 由于供应严重受限,三星和SK海力士主导市场条款。大型科技超大规模云服务商(微软、Meta、Google、Amazon)陷入AI军备竞赛。他们目前对价格高度不敏感,优先考虑确保供应量而非讨价还价。 因此,这看起来是一个明确的买入机会,尤其是鉴于预测范围: 摩根士丹利 | 麦格理对三星: 2026年:~1820亿美元 vs. ~2108亿美元 2027年:~2351亿美元 vs. ~3337亿美元 摩根士丹利 | 麦格理对SK海力士: 2026年:~1329亿美元 vs. ~1905亿美元 2027年:~1670亿美元 vs. ~3128亿美元 此次抛售看起来像是出于对2022年俄乌冲突的恐惧以及10倍杠杆工具带来的极端杠杆进行的即时去风险化。 但与2022年不同,当时笔记本电脑和消费需求已经低迷,LNG飙升导致三星/SK海力士买单: 这次成本转嫁给了超大规模云服务商,因为需求太高。 人们总是将KOSPI比作白银,但区别在于前者拥有惊人的运营收入,在牛市情景下,SK海力士等公司两年的收入 alone 就会超过其市值。 三星/SK海力士的下跌看起来是一个明确的买入机会,因为他们的运营收入可能会轻松超越任何短期波动。

英文原文

Thoughts on the $EWY / KOSPI crash and whether it's a buying opportunity for Samsung/Sk Hynix: This looks like a clear buying opportunity. Especially as investors misunderstand that 2026 is not 2022 (Ukraine-Russia) conflicts. AI has fundamentally changed the demand for memory, compared to the 2022 downturn in consumer electronics, where Samsung/Sk Hynix also footed opex costs. The main catalyst for one of the largest single-day drops in Korean history was Iran drone striking QatarEnergy's Ras Laffan complex. Qatar is roughly ~20% of the global LNG supply. Results: -> Dutch TTF (Euro Natural Gas): +46% -> Asian Spot LNG: + 39% -> U.S. Natural Gas: +4.16% (fine) The LNG spike was the catalyst for the drop as South Korea is materially exposed as it relies on imported LNG. Fearing this inflationary energy shock, investors aggressively dumped South Korean tech equities, combined with deleveraging of 10x Samsung/Sk Hynix. Sending -> Samsung Electronics down by 9.79%. -> SK Hynix shares sank by 11.12%. and the $EWY (KR futures) down 14.49% today. Everyone is fearing the 2022 incident with Russia-Ukraine war increasing LNG prices, causing industrial electricity tariffs in South Korea climbed by roughly 70% after KEPCO stopped abosring losses. In 2022 alone, Samsung's domestic facilities consumed approximately 28,000 GWh of electricity. The tariff hikes added an estimated KRW 2 trillion (over $1.4 billion USD) to Samsung's annual operating expenses and over KRW 1 trillion for SK Hynix. This unprecedented spike in operational expenses (OPEX) hit at the exact worst possible time. In late 2022, the semiconductor industry entered a brutal cyclical downturn. Post-pandemic demand for PCs and smartphones plummeted, leading to a massive oversupply of DRAM and NAND flash. However, 2026 is completely different: Compared to 2022 that faced increasing energy costs during a downturn: AI demand for memory is completely unprecedented. On the 27th, there were reports of DRAM hikes again from Samsung/Sk Hynix. Demand for NAND prices from $SNDK have been so high for capacity allocation that they've been taking 3 year pre-orders in advance. There is no end to demand in sight, and no relief way until 2028. Comparing this to 2022, where consumer segments for laptops and smartphones were already facing a downturn, memory makers could not pass on the costs (and tanked opex bill). That caused a major correction last time. Because the global AI chip shortage is so acute in 2026, they will seamlessly pass their inflated utility bills down the supply chain. Because supply is severely constrained, Samsung and SK Hynix dictate the market terms. Big Tech hyperscalers (Microsoft, Meta, Google, Amazon) are locked in an AI arms race. They are highly price-insensitive right now and prioritize securing volume over haggling. As a result, this looks like a clear buying opportunity, especially as projections range from: Morgan Stanley | Macquarie on Samsung 2026: ~$182.0B USD vs. ~$210.8B USD 2027: ~$235.1B USD vs. ~$333.7B USD Morgan Stanley | Macquarie on SK Hynix: 2026: ~$132.9B USD vs. ~$190.5B USD 2027: ~$167.0B USD vs. ~$312.8B USD The selloff looks like an immediate de-risking from fears over 2022 Ukraine-Russia-conflicts, combined with extreme leverage from 10x instruments. But unlike 2022, where laptops and consumer demand were already facing a downturn and LNG spikes caused Samsung/SK Hynix to foot the bill: This time it's passed onto hyperscaler costs because demand is too high. People always compare KOSPI to silver, but the difference is one has whopping operational income, where in a bull-case scenario, 2 years of income alone would exceed their market cap of companies like SK Hynix. The drop on Samsung/Sk Hynix looks like a clear buying opportunity as their operational income will likely blow past any short term volatility.

在 X 上查看原推 ↗