Yesterday's conversation also mentioned that I personally don't feel very good this week. When Huang Renxun said he wanted to develop AI PC, funds rushed into the PC and CPU industry chains; He said he wanted to promote the AI Factory, but funds flooded into areas such as liquid cooling, high-voltage electricity, power distribution, and electricity; Yesterday he said that Marvel would be a trillion dollar company, and on the same day, it rose by 30%. Its original market value of 200 billion dollars suddenly increased by tens of billions, which is actually a bit fanatical. Any parabolic rise in assets is unsustainable and there will always be adjustments. I didn't expect it to come just after we finished discussing adjustments yesterday The current market support for the entire logic is that the commercialization of AI has entered a turning point and is growing rapidly. Only with rapid commercial growth can the massive capital expenditure appear reasonable, and at the same time, the entire AI computing power industry chain upstream and downstream can continue to benefit. So the key is to see if the growth rate of AI commercialization has slowed down. 1) As long as the annualized revenue of large model manufacturers continues to grow and the big technology cloud business continues to exceed expectations, the entire business logic has not been reversed; At this point, if the short-term price rises too much, the price falls short of expectations, and the funds feel expensive, the market will retreat. If the market falls to a medium level, and there is a new opportunity later (such as the financial reporting season in April this year, coupled with the rapid growth of annualized revenue), the market will rise again. 2) If the performance of future big model manufacturers falls short of expectations, it is equivalent to a problem at the bottom of the market narrative logic, because Microsoft Google、 Many of Amazon's computing power purchases come from large model manufacturers, who are more upstream and closer to the origin of commercialization. In this case, it will at least be a moderate adjustment, and all logic needs to be reset, not completely starting over, but a moderate reset. Then everyone will go back and wait: you need to provide new evidence to prove that it has returned to rapid growth, and its scale and growth rate have exceeded expectations before confidence can return.