Post

三星跌完海力士跌，但这次真不一样今天早上，韩国股市又崩了。韩国KOSPI指数一度跌超5%，触发熔断。SK海力士跌超10%，三星电子跌超6%。一个月，SK海力士跌了35%，三星跌了23%。你身边有没有这种朋友——年初追高韩国芯片股，喊着“AI时代内存为王”，现在看着账户，整个人都不好了？别急着割。高盛今天发了一份报告，说了一句话： “韩国股市已经抛售过度。” 他们把KOSPI未来12个月目标价定在12000点。逻辑很简单——内存周期可能比以往更强、持续时间更长，而现在的股价根本没反映出这一点。高盛还重申三星电子和SK海力士“买入”评级，目标价分别看49万韩元和350万韩元。你可能觉得高盛在唱多出货。但看两组数据：第一，库存。截至二季度末，三星和海力士的DRAM及NAND库存只有2到4周。正常水平是4到5周。历史上每次下行周期前，库存都在10周以上。现在的库存，连正常水平都不到。第二，产能。三大原厂2027年的DRAM和HBM产能，全部卖光了。客户最终拿到的货，只有他们最初请求的60%到70%。 2027年，业内已经在喊“存储最缺一年”。更狠的是一件事—— 苹果想去长鑫存储压价，被拒了。苹果，全球最强势的采购方，过去谁能拒绝苹果的降价要求？长鑫存储直接说：报价不能低于三星和海力士。连备胎都没有了。 DRAM的议价权，已经从终端厂商完全转到了存储企业手里。这不是周期。这是结构性短缺。那问题来了：内存供给持续紧张，谁才是真正受益的？芯片公司？三星和海力士在涨价——7月DRAM合约价环比涨14.3%，创历史新高。但股价在跌。因为市场担心“AI资本开支见顶”，把芯片股当成了AI周期的衍生品一起砸。设备公司？东京电子、ASML这些卖铲子的，订单排到2028年。但设备公司的估值已经打满了预期，再涨需要超预期的东西。还是AI产业链？ HBM全部被英伟达包圆，三大原厂2027年HBM产能售罄。但AI芯片公司自己的股价也在调整。我的看法很简单—— 这一轮内存周期，真正的受益者不是某一个环节，而是整个“存储”这个资产类别本身。 DRAM正在从“配件”变成“战略物资”。以前内存是 commodity，价格周期三年一轮。现在内存是 bottleneck，谁有货谁说了算。苹果压不下价、产能提前两年卖光、库存只剩两周——这不是周期顶部该有的样子。最后说句实在话：市场把韩国芯片股当AI概念股在砸，但内存的基本面跟AI叙事是两个东西。 AI可以降温，但服务器要内存、手机要内存、汽车要内存。DRAM的需求是刚性的，供给是垄断的。三星和海力士一个月跌了23%和35%。高盛说这是过度抛售。你觉得是错杀，还是周期见顶？ $SKHYNIX $SKHY $SAMSUNG #内存卖方市场延续，韩股能否迎来反转？

Why should a public chain that hasn't even gone mainnet yet have BlackRock, Visa, and DTCC all act as its "security guards"? If there is a public chain that hasn't been on the mainnet, hasn't run real transactions, or tested its performance—would you invest? Most people won't. But BlackRock will. Visa meetings. DTCC will do so. Mastercard will do the same. On August 5th, Circle released a list: BlackRock, DTCC, Galaxy, Global Payments, ICE, Mastercard, MoneyGram, SBI Group, Standard Chartered Bank, Sumitomo Corporation, Visa. 11 companies. All are top stars in Wall Street and the global payments sector. A public chain that hasn't yet launched on the mainnet has gathered half of the world's financial infrastructure. This is not to "support." This is the role of a validator—the one responsible for cybersecurity. Think carefully about this logic: Generally speaking, validators in public chains are miners, stakers, and tech geeks. But Arc's validators are BlackRock, Visa, and DTCC. Institutions that rely on network integrity are also responsible for safeguarding cybersecurity. This is not "institutions coming to endorse"; it is "users are guardians." DTCC — The clearing hub for almost all securities trading in the United States. What does it mean to be a validator? This means Wall Street's settlement layer is moving toward Arc. Looking at the specific collaborations, they're even more explosive than the list itself: BlackRock: Deploy the BUIDL tokenized money market fund on Arc. Institutional investors can complete subscriptions, redemptions, and fund deployment on-chain in one place. Traditional fund subscriptions require T+2, form filling, and a bunch of procedures—all handled on Arc with a smart contract. DTCC: Plans to tokenize DTC-custodial assets on-chain in the second half of 2027, enabling native settlement with stablecoins. This is no small fry—DTCC handles securities liquidations worth hundreds of billions of dollars across the United States. BNY and Standard Chartered: Exploring integration of digital asset custody, foreign exchange, and repo infrastructure. This is not a "try and see." This is "bringing up the entire savings." Now, a detail. In May this year, Circle held a presale of Arc tokens, raising $222 million and valuing it at $3 billion. The A16z led the investment at $75 million. BlackRock, ICE, SBI, and Standard Chartered Ventures all followed suit. This isn't the first time these institutions have put in money. They paid twice—once as an investor, once as a validator. Some say: Isn't Arc just a permissioned blockchain? What about decentralization? But have you ever thought about it? What Wall Street really wants is not "decentralization," but "trusted centralization." Regulation must be compliant, institutions must be safe, and liquidations must be certain. Arc's validator model directly raises compliance standards to Wall Street standards. Ethereum addresses "no trustlessness." Arc addresses "trustworthiness." These two things serve different groups of people. Circle's Q2 earnings report just released: revenue of $701 million, USDC circulating supply of $73.3 billion, up 19% year-on-year. On-chain transaction volume reached $14.8 trillion, a year-on-year increase of 151%. USDC is still rising. But the narrative of USDC is about to change. Previously, USDC was a "stablecoin"—a payment tool and medium of exchange. After Arc launches, USDC will become the native currency of the Wall Street settlement layer. If USDC is Circle's present, then Arc is what Circle is trying to define—the next-generation financial settlement layer on Wall Street. On September 16, the Arc mainnet launched. Let's look at two things when the time comes: First, has BlackRock BUIDL really started running? Second, has DTCC's tokenization roadmap been progressing? If these two events happen—the story of the stablecoin sector will no longer be about "payment," but about "settlement." And USDC will be the native currency for that settlement track. A public chain that hasn't even gone mainnet has gathered half of Wall Street as a validator. This is not a technical narrative. This is a narrative of power. $BTC $CRCL $COIN #Circle财报后押注Arc, can USDC experience new growth?