Silver mining stocks lead semiconductor by four months? What institutions are truly concerned about is the decline of AI trading While reading the research report today, I happened to come across an institution comparing silver and semiconductor stocks. To be honest, I didn't take it seriously myself, just like before when I mentioned the association between M2 and Bitcoin. However, I happened to see that Mr. Mai also talked about this topic, so I talked about this data. This does not represent my personal opinion, I am only interpreting the data released by the institution. The reason why institutions compare silver mining stocks to semiconductors is not because there is a direct relationship between silver and chips, but because these two types of assets are traded by the market as "high elasticity cyclical assets" at certain stages. Silver mining stocks are very sensitive to silver prices. When silver prices rise, mining stocks rise even more fiercely, and when silver prices fall back, mining stocks fall even faster. In semiconductors, especially in storage HBM、DRAM、NAND， There is also a similar cyclical property, where when demand is strong, prices rise, and customers are scrambling for goods, the stock price elasticity is very large. Once the market begins to doubt the demand and price cycle, funds will quickly withdraw. So institutions are looking at the liquidity, risk appetite, commodity cycle, inventory cycle, and price cycle reflected behind silver mining stocks. Silver mining stocks typically react faster to these factors because they are inherently linked to commodity prices, real interest rates, the US dollar, and risk appetite. Semiconductor reactions may be slower because chips still need to go through processes such as orders, inventory, customer capital expenditures, and financial guidance, which can cause a lag in market confirmation. In human terms, if silver mining stocks rise first and then fall, institutions may think that this may indicate that the same type of high elasticity cycle trading has begun to decline, while semiconductors have not fully reflected it yet. This is what is called 'leading by 4 months'. It's not that silver has fallen, so semiconductors are bound to fall. What I really want to express is that silver mining stocks may reflect changes in capital's preference for cyclical assets, commodity prices, and risk in advance, while in this round of semiconductor market, especially stocks related to storage and AI infrastructure, it is becoming more and more like a cyclical trading cycle. The previous market trading is that major manufacturers continue to buy chips, computing power remains tight, storage prices continue to rise, and AI capital expenditures continue to accelerate. But now the market is asking another question, are these big companies buying too little, or have they already bought too much? When will AI revenue catch up with such a large capital expenditure? If this question is not answered correctly, the first part of semiconductors to be re priced is likely to be the most commodity like part, which is storage. Speaking of which, some institutions have already started to examine some semiconductor stocks in the form of cyclical stocks. Although I don't fully agree with this viewpoint, it is consistent with the US stock market data I recently wrote. The flow of funds in technology stocks is starting to deteriorate, with institutions and hedge funds reducing their holdings. Nearly 60% of technology stocks have fallen by more than 20% from their one-year highs. Now, some institutions are using silver mining stocks as an analogy to semiconductors, indicating that AI trading is no longer as FOMO as it was a while ago. Previously, as long as we talked about AI, computing power, and chips, the market was willing to give a premium. Now, the market is demanding verification, demanding to see orders, cash flow, capital expenditures, and profit returns, which means the importance of financial reports is becoming more detailed. My understanding is that the long-term logic of AI is still there, but in the short term, semiconductors may have already entered the verification stage.