Post

财报一个比一个炸裂，股价一个比一个崩。闪迪Q4赚麻了，股价膝盖斩；海力士HBM卖断货，单日暴跌19%；美光FQ4指引500亿，7月照样被锤28%。这到底是黄金坑，还是周期顶？我倾向于——坑的可能性还没被证实，顶的信号倒是一个接一个。三个信号对着看。第一，海力士自己先松口了。 8月9号韩国那边传出来，海力士在HBM上率先放缓涨价节奏，开始以低于同行的价格跟客户谈。大摩还在喊目标价370万韩元、说2026年股东回报超100万亿韩元，但市场闻到的是另一个味道：连HBM这种全行业最紧缺、毛利率最高的品种，涨价都开始乏力了。再叠加黄仁勋把Rubin Ultra的HBM配置从12层384GB砍到8层192GB、全球HBM需求预期被砍掉约10%，海力士刚经历韩股盘中-19%的历史最大单日跌幅，308万亿韩元市值蒸发。客户在砍、价格在松，你说牛市完好无损？我不信。第二，美光的独立行情能撑多久？美光这波确实硬，FQ3实际营收干到414.6亿、毛利率84.9%，FQ4指引直接给到500亿、86%毛利率、EPS 31，年化EPS run-rate已经120刀往上，HBM4量产、2026产能卖光，数据中心收入年化run-rate破千亿，逻辑硬。但HBM总共就三家能做，海力士都开始让价了，美光能独善其身？HBM是寡头市场，一家松口就是价格信号。而且美光7月也被锤过28.7%，现在的反弹更像超跌修复，不是新高起点。第三，$SNDK 闪迪自己的账。 Q4营收89.7亿/+372%、EPS 39.25、毛利率84.6%历史新高、939亿积压订单、140亿回购——放任何板块都是炸裂。但市场不认，因为Q1指引103-108亿，中值105.5，华尔街要111.5；毛利率指引83-85%，市场预期86.7%。杰富瑞更直接，目标价从3000大刀砍到1750，理由是"短期盈利增长最快的阶段可能正在过去"。从2354跌到1184膝盖斩，反弹到1250连下跌趋势线都没摸到。 $BTC 呢？还在6万5装死。别笑，真的是情绪指标。BTC 6万5附近趴着，从最高点到现在也膝盖斩了。AI存储、HBM、加密本质是这轮流动性+科技叙事堆出来的高Beta资产。BTC横两个月不创新高，说明场外增量资金没进场。这种背景下存储想靠超跌反弹反转，缺底气。反过来，如果8月13号闪迪真把情绪点着了，美光海力士跟涨，BTC这种高Beta大概率也会被带一下——同一条绳上的。所以AI内存牛市还稳不稳？我的判断：长逻辑没破，短周期见顶信号在累积。长逻辑是AI对存储的需求是结构性的——HBM4、DDR5、企业级SSD、NAND长协，不是炒概念，是真有云厂商排队拿货。但短周期的问题是：最好的预期已经打满，毛利率84%还能往哪涨？HBM都松价了，NAND拐点还远吗？海力士被砍单是第一道裂缝，闪迪指引不及预期是第二道，下一个是谁？ 8月13号，验真的日子。闪迪CEO和CFO出来答辩，市场要的不是Q4赚多少，是NBM长协结构、对海力士砍单的回应、2027年NAND市场拆分。拿出硬数据，1200是黄金坑，美光海力士BTC都能喘口气；打太极讲愿景，这反弹就是逃命波，还得下台阶。财报炸裂但股价跌、同行松价、BTC躺平——你觉得这是牛市中继，还是周期在敲门？$BTC #存储股抛压缓和，AI内存牛市还稳吗？

Countdown to SanDisk Investor Day: Why does the market reject this explosive earnings report? SanDisk's Q4 earnings report might be hard to believe—revenue of 8.97 billion, a year-over-year surge of 372%; adjusted EPS of $39.25, compared to just $0.29 in the same period last year, a 135-fold increase; gross margin at 84.6%, the highest in history. They also approved a $14 billion buyback, hold 8 long-term NBM contracts, and have backlog orders totaling 93.9 billion. With these results, the stock price dropped 7% in after-hours trading right after the report, then plunged as much as 14% the next day, falling from the June high of $2354 to a low of $1243—almost halving. Isn't it surreal? Where's the problem? The guidance. SanDisk's revenue guidance for next quarter is 10.3–10.8 billion, midpoint 10.55 billion. What does Wall Street expect? 11.15 billion. That's a full 600 million short. Gross margin guidance is 83–85%, market expectation is 86.7%, basically flat or slightly down quarter-over-quarter—this is a signal in cyclical stocks: the most profitable times may be behind. So now there's a huge divide between bulls and bears: The bulls' logic is strong: 93.9 billion in long-term contracts locks in future revenue, AI storage demand is still booming, data center business is growing several times over, NAND market management expects to reach a 500 billion scale by 2027. The $14 billion buyback is on the table, valuation has halved from the peak, PE is only 16x, looks like a bargain. The bears' logic is also solid: Your guidance has been "beat expectations but not enough to beat expectations" for two consecutive quarters, what does that mean? It means the best expectations are already priced in, future growth must continuously exceed expectations to support the stock price, and once growth slows, it's a double whammy. With an 84% gross margin, how much more upside is there? Insiders are still selling. Not to mention NAND is inherently cyclical—when it rises, it's fierce; when it falls, it's deadly. August 13 Investor Day is when SanDisk management will "defend" their position. CEO Goeckeler and CFO Visoso will personally present, expected to discuss AI storage long-term strategy, product roadmap, and how NBM long-term contracts smooth out cyclical fluctuations. Simply put, the market doesn't want to hear how much you earned in Q4, but whether you can prove: this time is different, not just another NAND cycle peak. My view: This stock is a classic "good company vs. good price" battle right now. The fundamentals are solid, but the stock price dropping from 2300 to 1200 is not without reason—the market is pricing in the possibility of a "cycle peak." If Investor Day can deliver convincing long-term growth logic (like AI storage penetration, contract coverage, capacity planning), then the current price might be a golden opportunity; but if it's just empty promises without concrete numbers, even 1200 might not hold. Stocks with good earnings but falling prices are the most tempting and the most dangerous. Do you think this is a mispricing or a justified drop? #闪迪8月13日投资者日临近，财报分歧待解