Bank of America: Seven Ongoing Risks Plague AI and Semiconductor Sector Stocks

Bank of America’s report flags seven persistent risks for stocks in the AI and semiconductor sectors, detailed below: 1. Hardware trading has turned into a "tourist trade" — even medical tour groups are discussing SNDK. Late buyers of MU may still face a washout, while sharp daily volatility is forcing committed bulls to exit positions. 2. Good news no longer moves markets. TSM, ASML, and INTC have declined even after reporting strong earnings, with intraday rebounds consistently met by selling pressure. Capital expenditure cuts by large-scale players will harm AI suppliers, while hikes will pressure free cash flow; financing activities are also stoking funding concerns. Ordinary upside surprises may no longer be enough to drive gains. 3. Investors are bearish on AI labs but still bullish on their infrastructure. Spending by OpenAI and Anthropic is already embedded in suppliers’ backlogs and financial data. Open-source weighted models may eventually boost 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 will also compress lab profits and intensify hardware and power optimization efforts. More compute power does not automatically translate to higher profits for every supplier. 5. Memory sector estimates may need to be revised downward. Lower valuation multiples alone may not suffice; stocks that appear cheap based on unrevised projections may not actually be undervalued. 6. Credit has entered the stock market debate. Financing for the entire ecosystem has become a "trillion-dollar game of chicken". 7. The sector’s ongoing correction may stem from position and factor adjustments rather than a single fundamental breakdown. Investors will later apply narratives around open-source models, ROI, financing, and optimization to price movements.