The market has once again committed an old flaw: treating a company's action as a judgment for the entire industry. Meta announced the rental of some excess computing power, and as soon as the news came out, the storage chip sector collapsed directly. Micron fell 10.57%, SanDisk fell 10.62%, Philadelphia Semiconductor Index fell 6.27%, Hynix fell 8%, and Korean stocks even triggered circuit breakers. It looks like ironclad evidence of the peak of AI capital expenditures and excess storage demand. But is it really like this? 2/First, clarify what Meta is doing. Renting out excess computing power does not mean that there is no demand for AI. A more likely scenario is that Meta heavily hoarded GPUs and computing power clusters in the first two years in order to train and infer large models. The training phase has come to an end, and some of the production capacity is temporarily idle. Renting it out is to improve utilization and recover costs. This is a capital expenditure rhythm adjustment for a single company, not a reversal of industry demand. The reason why the market has reacted so strongly is because AI related stocks have risen too quickly in the past two years, and valuations have been filled with too many optimistic expectations. Any slight movement will be interpreted as the peak of the cycle. However, in the field of storage, the cycle of traditional semiconductors cannot be simply applied. AI training requires HBM, AI inference requires DDR5 and high bandwidth memory, and data center expansion requires SSD. These demands are not a one-time deal, but a structural growth that has persisted for many years. More importantly, the real supply bottleneck has not been resolved yet. HBM is still dominated by SK Hynix, Samsung, and Micron, and capacity expansion requires huge capital expenditures and long-term technological accumulation. The yield, packaging, and testing of high-end storage are barriers at every stage. Even if Meta has surplus computing power in the short term, the demand from other cloud vendors, enterprise customers, and sovereign AI projects around the world is still in line. With the release of Nvidia's new chips, the demand for HBM will only increase, not decrease. From a historical perspective, the stock price of storage chips has always been like this: when it rises, it talks about a demand revolution, and when it falls, it talks about a cyclical peak. But after every major drop, companies that truly control high-end production capacity will reach new highs. In 2019, memory experienced a significant drop, followed by the outbreak of the pandemic and data centers, causing Micron to multiply several times. The storage winter of 2022, AI has pulled it back in 2023. This time is no exception. The only difference is that this time the demand driving force has been replaced by AI, and the narrative of AI has just begun. So my judgment is: This is not the top of the storage industry, but a buying window brought about by an emotional overreaction. Companies like Micron and Hynix hold the rarest production capacity in the AI era, and their long-term logic remains unchanged. Short term stock prices may continue to fluctuate, but for long-term investors, this panicked decline is actually an opportunity to increase their holdings. 7/Finally, let me say: Investing in storage is essentially a gamble that 'data won't get cheaper'. The more AI develops, the more data, the more expensive computing power, and the more important storage becomes. The market panic will pass, but data growth will not stop.