#我的七夕交易分享 AI demand is reshaping the value proposition of every layer of the supply chain, but the pace and certainty vary greatly across layers.
In the short term, consumer NAND is moving sideways while enterprise NAND is still rising, with Kioxia allocating 60% to 70% of its capacity to enterprise products. Eight 64GB memory modules now cost close to the price of an entire server, while enterprise rack-server shipments are shrinking. Supernodes will reach 7,000 racks in 2026, with mass production beginning in October; a full rack with 64 cards will cost 10 million to 11 million. Indium phosphide will face a clear shortfall in 2026, red phosphorus prices have tripled, and MOCVD lead times are 15 to 20 months. Meta will have 7GW of computing capacity by year-end, and the Personal Super Intelligence Agent may be launched in September. Toyoda weaving machines have a 1.5-year lead time, the electronic-fabric shortfall is one-third, and new kilns coming online in September will release monthly capacity of 4 million meters.
In the medium term, SanDisk's NBM has locked in $93 billion in revenue over three years; HBF samples will be delivered in 2027 and ramp up in 2028, with a maximum option value of $330 billion. Meta will reach 16GW by the end of 2027, split evenly between training and inference. Supernodes will reach 22,000 racks in 2027, with penetration of 35% to 38%, while 224G backplane connectors will begin ramping up. Domestic AI chip demand will reach 4 million to 5 million units in 2027, but 7nm process capacity is the bottleneck. The risk of a consumer NAND shortage will rise in 2027; materials bought for 1 billion in the past now cost 16 billion. Indium phosphide supply will exceed demand in the second half of 2027, and prices may fall. The electronic-fabric shortage will initially ease by the end of June 2027, while the trajectory of low-end fabric prices will depend on whether domestic weaving machines can achieve a breakthrough. Demand for high-end electronic fabric will rise to several times its current level in Q1 to Q2 2027, and the shortfall will persist.
In the long term, NAND will be revalued from a cyclical product into AI-inference storage. HBF places Flash inside the GPU package, offering 8 to 16 times the capacity of HBM at one-fifth to one-eighth the price. The baseline HBF TAM in 2030 is $47 billion. The inference market will be highly diversified, diluting Nvidia's share. Domestic chips are entering the elimination phase. SanDisk's $100 billion base market capitalization is the floor, while successful HBF validation will open a second growth curve. The next stage will focus on four things: whether HBF samples will be validated by cloud providers in 2027; whether Meta will launch its superintelligent Agent at Connect in September; whether indium phosphide supply will exceed demand as scheduled in Q4 2027; and whether domestic weaving machines can achieve a breakthrough in 2026, which will determine the trajectory of electronic fabric in 2027. $SNDK ‌

#股票交易分享挑战 SanDisk's 84.6% gross margin: A storage supercycle, or the final frenzy?
SanDisk's earnings report is indeed somewhat astonishing. Revenue for the latest quarter was $8.97 billion, up 51% QoQ; net profit was $6.9 billion. Even more strikingly, gross margin reached 84.6%, while revenue guidance for the next quarter remains at $10.3 billion–$10.8 billion.
The question is: for a company making NAND flash memory, with a gross margin already exceeding that of many software companies, does this mean AI storage is just getting started, or that industry profits are already nearing a peak?
First, understand where the profits come from
SanDisk itself provided the answer: approximately one-third of the quarterly revenue growth came from higher shipment volumes, and two-thirds came from price increases. This is important. It shows that SanDisk's earnings explosion is not just because more AI servers are being sold; more importantly, NAND supply is tight, giving the company strong pricing power. Data center revenue reached $2.98 billion, doubling QoQ, also proving that AI is gradually moving from “buying only GPUs and HBM” toward large-scale procurement of enterprise SSDs.
Therefore, the storage upcycle is not just a story; the profits have already been realized.
How is this different from traditional cycles?
Past storage cycles were simple: prices rose—manufacturers expanded capacity—oversupply emerged—prices collapsed. SanDisk is now trying to break this cycle. The company has signed new types of long-term agreements with eight customers, covering approximately 50% of shipments in fiscal 2027 and about two-thirds in fiscal 2028. The agreements include committed purchase volumes, pricing mechanisms, and minimum financial guarantees.
Put simply, customers cannot cancel orders at any time just because they see prices falling. This will reduce volatility in the storage industry and gives SanDisk the confidence to set a long-term gross margin target of approximately 80% for fiscal 2028–2030.
However, this is still only the company's target, not an achieved fact.
Who in the A-shares market truly benefits?
The easiest mistake here is to see SanDisk making huge profits and then buy every “storage concept stock” indiscriminately. There is currently no NAND manufacturer in the A-shares market that directly corresponds to SanDisk. Jiangbolong, BIWIN Storage, and Demei Li mainly purchase storage wafers and then sell controllers, packaging, firmware, and modules. They do benefit, but the logic is not exactly the same.
In the early stages of a storage price increase, the low-cost inventory held by these companies can generate significant earnings leverage. Jiangbolong expects first-half net profit of RMB 9.2 billion–RMB 11 billion, while BIWIN Storage expects RMB 7 billion–RMB 7.5 billion; their results have already been clearly realized. But if prices rise too quickly, subsequent procurement costs will also increase. If end customers do not accept further price increases, module manufacturers' gross margins could come under pressure instead.
Therefore, do not look only at net profit growth of several thousand percent; also examine inventory turnover, operating cash flow, and the proportion of self-developed controllers.
My conclusion: I remain bullish on the storage upcycle, but this is no longer the stage of “buying storage stocks with your eyes closed.”
SanDisk's earnings report tells us that AI data centers are becoming a new demand center for NAND; it also reminds us that a significant portion of the current profit growth comes from price increases. The industry trend remains upward, but whether stock prices can continue rising will depend on whether volume growth can take over from price increases.
Among A-shares companies, those whose results have already been realized and that possess self-developed controller and enterprise-grade product capabilities are more worthy of tracking; companies that rely solely on the appreciation of low-cost inventory have substantial earnings leverage, but equally substantial risks.
From here, focus on just three indicators: NAND prices, enterprise SSD shipment volumes, and storage companies' operating cash flow. If prices stabilize at high levels and shipment volumes continue to grow, this cycle can continue; if prices rise while shipment volumes begin to decline, then an 84.6% gross margin may not be the starting point, but rather a high point that warrants caution.$SNDK