Why Micron Stock Just Dropped Again

Micron (MU) fell 7.7% by 2:50 p.m. ET amid Warren Buffett’s warning that investors are increasingly “preferring gambling” in AI-driven themes. The article argues the stock still looks cheap at ~22x earnings versus analysts forecasting a ~10x earnings improvement this year, with profits doubling next year. Overall, price action is being driven by swing trading around semiconductor news, but the fundamental valuation case suggests limited downside despite the near-term volatility.

Analysis

Buffett's 'gambling' framing matters because MU is being traded as a momentum proxy, not as a cash-flow compounder. In that setup, the stock's multiple is more sensitive to changes in investor attention than to the next quarter's EPS, so any sign of slower AI capex or weaker memory pricing can trigger a faster de-rate than fundamentals alone justify. By contrast, NVDA still sits one step upstream in the AI budget chain and has more pricing power, so crowding risk is lower there than in commodity memory names.

The second-order risk is supply discipline breaking down. If SKHY accelerates wafer starts, the market will price in a longer period of excess supply even before spot prices turn, which is why MU can underperform on 'good' demand news: the key variable is not demand growth, it's the duration of scarcity. That makes this a 1-3 month catalyst trade around pricing checks and next guidance, not a 6-18 month structural short unless inventories start rising again.

Contrarian view: the move may be too reflexive. MU already screens cheap on forward earnings, so a skeptical tape can create a better entry if the next print confirms HBM/DRAM tightness; the setup flips quickly if gross margin or ASP commentary remains firm. The clean falsifier is a stable or accelerating memory pricing tape into earnings; if that happens, the current selloff is sentiment noise rather than a thesis change.

Key Decisions for Investors

• Short MU on rallies or buy a 1-3 month put spread into the next earnings window; the edge is multiple compression from crowded positioning, with upside to the trade if consensus revisions start rolling over. Falsifier: DRAM/HBM pricing stays firm and MU lifts gross margin guidance.
• Pair long NVDA / short MU over the next 4-8 weeks. If AI spend remains concentrated in compute, NVDA keeps its scarcity premium while MU remains the higher-beta victim of inventory fear. Cut the pair if MU starts outperforming NVDA for several sessions or memory pricing inflects higher.
• Use SKHY as the confirmation signal: add to MU short only if supply growth commentary points to more wafer starts or capacity expansion. If SKHY signals discipline, cover the short because the market will unwind the oversupply narrative quickly.
• No fresh long in ON until the market stops penalizing semis on flow rather than fundamentals. This is a timing issue, not a value call; wait for a cleaner earnings/revision catalyst before putting capital to work.