Post

挖矿的小羊
挖矿的小羊
Show original
闪迪一夜暴涨6000亿,但我关心的只有一件事:这饼能吃到2028吗? 8月13日,闪迪办了场投资者日。 这不是普通的财报电话会,这是公司从西部数据拆分后最重要的一次战略沟通。 为什么重要?因为闪迪的股价从6月2354美元的历史高位,已经跌掉了近一半。 市场在问一个问题:AI存储的故事,还能讲多久? 闪迪的答案,直接把华尔街干懵了。 先看数字: 2028到2030财年,营收保持中高双位数增长。 非GAAP毛利率维持在约80%。 非GAAP营业利润率约75%。 调整后自由现金流利润率约50%。 运营费用只占营收的5%。 更狠的是——完成业务投资后,100%的超额现金返还给股东。 高盛直接给了2200美元的目标价,说还能涨44%。 一个存储芯片公司,毛利率要做到80%。 你知道苹果毛利率多少吗?46%左右。 英伟达毛利率也就70%出头。 闪迪说:我要做到80%。 但等等——故事讲得再漂亮,饼画得再大,能不能吃到嘴里是另一回事。 闪迪自己也知道市场在担心什么。 存储行业有个臭名昭著的特点:周期性强到离谱。 涨的时候涨上天,跌的时候跌到妈都不认识。 过去一年,闪迪的股价从40美元左右飙到2300美元以上,然后又暴跌近一半。 这种过山车,谁敢长期持有? 所以闪迪这次拿出了一个大招——NBM长期客户协议。 说白了就是:客户提前锁单,我提前锁定产能,大家一起把周期熨平。 目前闪迪已经和8家核心客户签了协议,其中包括3家美国头部超大规模云服务商。 这些协议覆盖了2027财年约50%的存储产能,2028财年三分之二的产能。 合同总价值约940亿美元。 就算按合同底价算,毛利率也能做到80%。 这就是闪迪敢喊出80%毛利率的底气——三分之二的产能已经被锁死了,价格也锁死了。 再说技术面。 闪迪正在押注一个叫HBF(高带宽闪存) 的新技术。 简单说,这东西同时具备高带宽内存的高速读写能力和传统闪存的大容量优势。 专门解决AI推理阶段的内存瓶颈。 8月初,闪迪刚和SK海力士一起发布了HBF的首个标准规范。 首颗HBF存储芯片已经完成流片,预计明年就能交付初始样品。 这是AI存储的下一个战场。 高盛说HBF的技术路线图给闪迪带来了“巨大的上行潜力”。 但——我必须说个“但是”。 高盛自己都承认:NBM协议能不能真正平滑行业周期,还需要时间验证。 协议是签了,客户会不会违约?价格下限能不能兜住?市场需求万一掉头怎么办? 这些都是问号。 闪迪的股价今年已经涨了超过600%。 任何风吹草动,都可能是30%的回调。 最后说几句真心话—— 这次投资者日,闪迪给了市场三样东西: 第一,一个漂亮的长期故事(80%毛利率、100%现金返还)。 第二,一个可信的落地机制(NBM协议锁定了三分之二产能)。 第三,一个未来的增长引擎(HBF技术)。 但故事终究是故事,饼终究是饼。 真正值得跟踪的,只有三个指标: NBM协议的履约率——客户有没有真的按承诺采购。 HBF的商业化进度——明年能不能准时交付样品。 NAND市场的供需变化——周期有没有真的被熨平。 这三个答案,至少要等到2027年才能看到第一批数据。 在那之前,所有的大涨大跌,都是市场在情绪里游泳。 最后一句: 闪迪画了一张2028年才能兑现的饼。 在那之前,你要想清楚—— 你是相信这个故事的人, 还是等饼熟了再吃的人? $SNDK $SKHYNIX $SKHY #闪迪投资者日后,长期目标成焦点
挖矿的小羊
挖矿的小羊
SanDisk surged 17% painting a big picture, but I advise you not to rush in Last night, did you see anyone like this around you? Watching SanDisk's stock price surge 17% intraday, feeling excited, and opening the trading app to chase. "AI storage demand is exploding, data center revenue surged 645%, gross margin 80%—isn't this the next Nvidia?" And then? They might have already bought at the peak. Don't rush. The story isn't that simple. First, let's talk about what SanDisk actually said this time. On August 13 Investor Day, SanDisk released a bunch of "explosive" long-term targets: Fiscal years 2028 to 2030— Revenue to maintain mid-to-high double-digit growth Non-GAAP gross margin about 80% Operating margin about 75% Adjusted free cash flow margin about 50% 100% excess cash returned to shareholders Goldman Sachs immediately gave a $2200 price target, saying it could still rise 44%. The stock price once rose over 17%, closing up nearly 14%. Sounds great, right? But did you know—just a week ago, SanDisk released an equally "explosive" earnings report. Quarterly revenue $8.965 billion, up 372% year-over-year. Data center business revenue $1.467 billion, up 645% year-over-year. Non-GAAP gross margin 84.6%. And then? After the earnings release, the stock price dropped 12% in two days. From the June all-time high of $2354, it fell to $1238, nearly halving. The better the performance, the worse the drop. Does this scene look familiar? Google, Tesla, SanDisk—"good earnings backfire" has become the biggest market trap in 2026. Why? Because the market always trades expectations, not facts. SanDisk fell 47% from the June high, not because the company got worse. It's because it rose too much—up 430% year-to-date. What does this mean? The market has already overdrawn the "AI storage super cycle" story in advance. Now you tell me "future can still grow mid-to-high double digits, gross margin 80%"—the market will only ask one question: "Can it be better than expected?" If not, then the current price is the ceiling. Even more painful—SanDisk is doing something the storage industry has never done before: fighting the cycle. NAND storage is a typical cyclical industry. Demand exceeds supply → price rises → capacity expansion → oversupply → crash → production cuts → demand exceeds supply again. This cycle hasn't changed in the past twenty years. SanDisk says: "I signed long-term agreements (NBM) with 8 customers, covering about two-thirds of shipments in fiscal 2028, with a total contract value of $94 billion. I want to smooth the cycle." Sounds sexy, right? But every company in history that tried to "smooth the cycle" was harshly taught by the cycle. The cycle can't be eliminated by a few contracts. When demand really turns down, contracts are just paper. So is AI storage still worth watching? Yes. But not at the current price. SanDisk's fundamentals are solid—AI inference is turning data centers into "storage-intensive" scenarios. By 2030, enterprise data center flash market is expected to reach 1.2ZB. HBF high-bandwidth flash technology is also accelerating. The long-term logic is strong, short-term valuation is expensive. These two things coexist without contradiction. 07. A few hard truths for crypto players: First, don't chase the highs. Those who rushed in when SanDisk surged 17% will likely lose out. Good company ≠ good price. Second, strength in the storage sector will spill over to the crypto market. SanDisk, Micron, Western Digital all surged, indicating the AI hardware narrative is still alive. Capital will look for the next low point—AI-related crypto projects, especially decentralized storage tracks, may be rotated into. Third, but don't bet on earnings. SanDisk has already shown you with experience—the better the performance, the higher the expectations; the higher the expectations, the easier it is to "fully priced good news." Last sentence: The market rewards not those who are right, but those who are right at the right price. SanDisk painted a big picture, very appetizing. But don't bite when the pie is hottest. $SNDK $SKHY $WDC #闪迪投资者日后,长期目标成焦点

Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more

Replies

No comments yet. Be the first to reply!