Samsung's Q2 earnings forecast is based on supermarket expectations, with revenue only slightly lower than the market's most optimistic expectations but far exceeding average expectations. Why is it still plummeting? 1) It's not that the performance is not good enough, but that the market is ahead of the performance. It is a typical logic of 'selling facts' - investors have already fully priced their extremely high expectations. When the valuation is stretched and the rotation is fully unfolded, even strong performance is not enough to meet the market. 2) Also mentioned in yesterday's tweet: "The key point of this season's earnings season is that expectations are already high and the margin of error has decreased. Expectations are very high, which means that the market not only wants companies to beat, but also companies to raise significantly. Speaking of the storage sector, did we talk about https://(x.com)/qinbufark/status/205417474022216464 in May? S=46&t=k6rimWSEbo2D2TXolYcM-A According to the expansion pace of several existing storage manufacturers, the new production line may not be put into operation until the second half of the 27th year, and the large-scale release of production capacity will be in the 28th year. In the coming year, storage will still be in a state of scarcity. It is also the current state of the market, with high levels, high leverage, and crowding. It is also the earnings season, and everyone knows that the market expectations are high. They are also worried about which earnings report will drop a bomb, so there is a risk aversion. The core is still what was discussed in yesterday's tweet: If Google Cloud aws、 Microsoft Cloud and other major cloud providers continue to exceed expectations, so the market will indeed continue to accelerate the commercialization of AI after a turning point in the first quarter. At the same time, several CSPs also need to confirm again that capex has not slowed down. Of course, the latest annualized ARR data officially disclosed by the two major model manufacturers is also a crucial core influencing variable, depending on the growth rate. A thought from yesterday to today is https://(x.com)/qinbafrank/status/207430906584984813? S=46&t=k6rimWSEbo2D2TXolYcM-A, CSP performance will be a key indicator in the commercialization process of AI, with weights equal to or even higher than those of closed source large model manufacturers in the future. Due to the adoption of AI process cost engineering by enterprises, many high-frequency and medium-term value scenarios will shift towards high cost-effective tokens. This part of the commercial value is actually cut from the closed source big model and will ultimately settle in the performance of CSP. CSP (Cloud Service Provider) will become the most comprehensive carrier to showcase the commercialization process of AI.