BlockBeats News, July 30th, Bank of America released a report stating that there are seven key ongoing risks in the AI and semiconductor industry stocks, as follows:

1. Hardware has become a "tourist transaction," with even a medical bus tour discussing SNDK. Those who are late to buy MU may still need to undergo a washout, while intense intraday volatility is also forcing strong bulls to sell.

2. Good news no longer works. TSM, ASML, and INTC are experiencing price declines even after delivering good performance, with intraday rebounds being consistently sold off. Large-scale capital expenditure reductions by major manufacturers will harm AI suppliers, while increases will pressure free cash flow, and financing will raise funding concerns. Being merely better than expected may no longer suffice.

3. Investors are bearish on AI labs but are still bullish on their infrastructure. The spending from OpenAI and Anthropic has been embedded in supplier backlog orders and financial data. The adoption of open-source weight models may eventually increase compute demand, but the transition process is unlikely to be smooth.

4. Token deflation is a double-edged sword. Cheaper tokens can accelerate adoption, but they can also compress lab margins and worsen hardware and power optimization. More compute power does not automatically mean more profit for every supplier.

5. Memory may need to be re-estimated. Simply lowering valuation multiples may not be sufficient. Stocks that appear cheap based on unrevised estimates may not actually be cheap.

6. Credit has entered the stock debate. Financing for the entire ecosystem has become a "tens of trillions of dollars cowardly game."

7. The ongoing adjustment of sector stocks may stem from holdings and factor adjustments, rather than a single fundamental breakdown. Investors will later apply the open-source model, ROI, financing, and optimization narrative to price trends.