Post

上半年累计营收首次突破100万亿韩元，同比增长257%。第二季度营收79.32万亿韩元，营业利润60.54万亿韩元，净利润同比暴增1242%，营业利润率高达76%。现金及现金等价物88万亿韩元（约616亿美元），环比暴增近62%，净现金头寸69.4万亿韩元。 HBM4已量产出货，HBM4E样品已交付客户。一季度HBM市场份额58% ，稳居全球第一。每一项数据单拿出来，都足够让任何一家公司开香槟庆祝。然后呢？ 7月29日财报发布当天，SK海力士韩股暴跌超17%，创历史最大单日跌幅纪录。美股上市以来，股价已从接近195美元的高位累计下跌约21%。若从6月高点算起，回撤幅度超过50%，市值近乎腰斩。利润最好的财报，成了股价最惨的催命符。市场到底在怕什么？怕的是存储行业那个著名的“死亡循环”—— 赚钱→扩产→产能过剩→价格战→亏损→破产。过去三十年，这个循环摧毁了无数存储公司。每一次行业景气高点，都是扩产狂潮的起点；每一次扩产狂潮，都是下一轮崩盘的伏笔。现在SK海力士在干什么？正在用历史上最猛的速度扩产。 2026年上半年，购置有形资产的资本支出达18.33万亿韩元，同比增长72.7%。全年资本支出预计达到40万亿至50万亿韩元区间高位。 SK集团更是宣布未来十年投资7200亿美元用于AI内存扩产，目标是将产能提升至当前的三倍。7月纳斯达克上市又募资265亿美元，全部砸向扩产。市场看到的是：一家公司把赚到的每一分钱，甚至借来的钱，全都扔进了新工厂。而存储行业的历史告诉市场：当所有人都觉得需求永远不够的时候，往往就是供给即将过剩的时候。但这一次，可能真的不一样。第一，长协锁死了需求。 SK海力士已与约10家核心客户签订长期供货协议，期限3到5年。其中与英伟达敲定5000亿美元合作方案。HBM的长期协议覆盖率甚至达到了100%。过去行业签一年长约，现在一签就是五年。五年意味着什么？意味着不管市场价格怎么波动，这些产能已经有人买单了。第二，供给真的跟不上。高盛测算，2026年全球DRAM供需缺口约5% ，2027年扩大至5.9% ，紧张态势将持续至2028年。摩根大通预测HBM供需缺口在2026至2028年分别约为-15%、-14%和-22%。 SK集团会长崔泰源的原话更直接：“所有的客户都在要求接近原来需求两倍的供货量，但供应完全跟不上。” 他甚至预测，2027年将是“存储芯片荒”最严重的一年。第三，EUV设备交期超两年半，产能不是说扩就能扩的。半导体工厂从动工到量产，至少需要四到五年。即便现在砸钱，新增产能也要到2027年底至2028年才能集中释放。需求的爆发是瞬时的，供给的响应是以年为单位的。市场不是在定价“海力士好不好”，而是在定价“这个行业会不会重蹈覆辙”。 557%的增长不够，因为市场怕的是——你现在赚的每一块钱，都是在为下一轮崩盘攒弹药。但如果你看得更远一点—— 长协锁定了五年需求、供需缺口持续到2028年、EUV设备扩产需要四五年、AI需求是结构性的而非周期性的—— 这一次的“死亡循环”，可能真的被打破了。崔泰源自己说：“与其说半导体行业周期消失，不如说周期被大幅拉长。” 这对加密市场意味着什么？第一，别用“利润增长”来判断一个资产的价值。市场定价的是“未来现金流能不能覆盖今天的资本开支”。海力士赚了557%，但把利润全砸进了新工厂——市场说“不够”。第二，长期逻辑和短期定价是两回事。海力士的长期逻辑无懈可击——AI需求、长协锁定、供需缺口。但短期市场看到的是：资本开支增速（72.7%）远高于利润增速（557%的基数效应掩盖不了边际放缓）。第三，加密市场里的“AI叙事”项目，正在经历同样的估值重构。当传统资本市场的AI硬件股都在被重新定价时，靠蹭AI概念拉盘的项目，只会跌得更快。最后一句： 557%的利润增长，在“死亡循环”的恐惧面前，只值一个跌停。但当长协锁死需求、供需缺口持续到2028年—— 今天的扩产，可能就是明天最深的护城河。 $BTC $SKHYNIX $SKHY #海力士扩产提速，资本开支能否兑现回报

How profitable is SK Hynix really? In Q2, revenue reached 79.3 trillion KRW (about $54.5 billion), a year-over-year surge of 257%. Operating profit was 60.5 trillion KRW (about $42 billion), soaring 557% year-over-year. Operating margin stands at 76.3%. For every 100 won of sales, they net 76 won. For the first half of the year, cumulative revenue surpassed 100 trillion KRW for the first time. Any one of these figures would make 99% of listed companies kneel and call them daddy. So what happened next? On the day the earnings report was released, the Korean stock market plunged over 19% intraday, marking the largest single-day drop in history. The KOSPI index triggered circuit breakers for two consecutive days. A record-breaking earnings report triggered a nationwide stock market crash. What is the market afraid of? It fears the death cycle ingrained in the storage chip industry: Price increase → capacity expansion → oversupply → crash → bankruptcy. This cycle has killed countless semiconductor companies over the past 30 years. In 2008, DRAM prices collapsed, and Germany's Qimonda went bankrupt. In 2015, during the storage chip winter, Micron lost over $2 billion. Every "supercycle" celebration ends in a mess. What is SK Hynix doing now? In the first half of the year, cash outflow for tangible asset purchases exceeded 18 trillion KRW, a year-over-year increase of over 70%. Full-year capital expenditure is expected to reach the high end of 40 to 50 trillion KRW. The first phase of the Yongin wafer fab has a total investment of 31 trillion KRW. The Cheongju M15X fab, M17 fab, and P&T7 packaging plant are all fully operational. SK Hynix is expanding capacity at the fastest pace in its history. The market watches all this with only one question in mind: "With so much capacity being added, who will pay the bill two years from now?" But this time, SK Hynix has learned its lesson. It didn’t blindly expand capacity to await death like in 2008. It did something no one in the storage chip industry has ever done—locking demand into contracts. In July 2026, NVIDIA and SK Hynix signed a long-term supply agreement potentially worth up to $500 billion. The agreement includes jointly developing next-generation AI memory and ensuring stable supply of HBM. This is not an isolated case. The market procurement model is undergoing structural change—large cloud service providers are shifting from traditional one-year contracts to multi-year supply agreements. Customers fulfill contracts with prepayments and deposits. Capacity is locked in advance until 2028, even 2029. What about the supply side? EUV lithography machine delivery cycles exceed two and a half years—want to expand capacity? Wait three years. Demand side locked, supply side constrained. This is how SK Hynix breaks the "death cycle." Bitcoin’s past three halvings each had people saying "supply reduction means price increase." What happened? After the 2022 halving, Bitcoin fell from 69,000 to 16,000. Why? Because there was only supply-side constraint, no demand-side lock-in. Miners can halve, but whales can dump. ETFs can buy in, but institutions can redeem. The biggest problem in the crypto market has never been "not scarce enough"—but "no one promises to keep buying." SK Hynix’s $720 billion (about 1,000 trillion KRW total investment) tells the world one thing: Surviving cycles relies not on faith, but on locking demand into contracts. Bitcoin has the "halving" as a hard supply-side constraint. But to this day, the crypto market still lacks a demand-side lock-in mechanism like "long-term agreements." ETFs are the first institutional-level long-term capital entry, but they can still be redeemed at any time. True "cycle survival" requires buyers and sellers to put volume and price for the coming years on paper, with penalties for breach. What does this mean for us crypto traders? First, stop treating "halving means price surge" as gospel. Supply halving doesn’t equal price doubling—halving without demand lock-in is just retelling the same story. Second, ETFs are the first step, not the last. When the crypto market sees "institutional-level long-term custody + non-redeemable lock-up mechanisms," that will be true cycle survival. Third, SK Hynix’s playbook tells us: truly smart money "locks demand" at cycle peaks, not "bets on price increases." The storage chip industry’s 30 years of blood and tears teach us— Those who die worst in cycles are always those who only believe in "supply scarcity" without locking in "demand certainty." The Bitcoin halving story has been told for over a decade. The next decade should tell some new stories. $SKHYNIX $SKHY $xSKHY #海力士扩产提速，资本开支能否兑现回报