ChainThink reports that, on July 30, according to a Bank of America report, AI and semiconductor industry stocks continue to face seven persistent risks, including crowded hardware trades, difficulty in driving stock prices with positive earnings, market skepticism regarding the profitability prospects of AI labs, deflationary effects of tokens on profits, potential downward revisions to memory expectations, increasing pressure from credit financing, and the possibility that sector adjustments may be driven by changes in holdings and factors.
Reports indicate that companies such as TSM, ASML, and INTC have faced persistent downward pressure on their stock prices despite delivering strong earnings, with intraday rebounds consistently met with selling pressure; if mega-cap manufacturers reduce their capital expenditures, it could impact AI suppliers, while an increase might suppress free cash flow and raise concerns about financing.
Bank of America also stated that spending by AI labs such as OpenAI and Anthropic has already been reflected in supplier backlogs and financial data.
Cheaper tokens may accelerate application adoption, but they also compress lab profits and drive hardware and power optimization—more computational power does not necessarily mean proportional profit increases for all providers.
The report suggests that memory-related stocks may not adequately reflect risk if they rely solely on downward valuation multiples; stocks that appear cheap based on unadjusted expectations may not actually be undervalued. The subsequent price movements of the AI and semiconductor sectors may continue to be influenced by financing, ROI, open-source models, and optimization narratives.