Micron's Broadcom‑Driven Drop Never Made Sense — Dip Buyers Are Up 40%

Micron stock has rebounded sharply, up 9.03% on Monday to $1,070.29 and roughly 42.6% over the past month, leaving it just below its $1,089.29 52-week high. The article argues the earlier selloff was misguided because Broadcom's slightly lower gross margin guide was driven by XPU mix, not HBM demand, while Broadcom still guided Q3 AI semiconductor revenue to about $5.1 billion, up 60% year over year. Micron’s HBM supply for 2026, including HBM4, is said to be sold out, reinforcing a constructive supply-demand backdrop.

Analysis

The key mispricing here is not about earnings quality, but about category error: investors briefly treated a memory supplier as if it were a custom-accelerator vendor. That matters because the AI capex complex is becoming more modular, and the bottleneck is shifting from compute silicon to memory and networking density; when inference deployments accelerate, HBM demand typically inflects with a lag but tends to be stickier than accelerator ASPs. In that setup, the second-order beneficiary is still the memory stack, while the more exposed names are the companies whose gross margin sensitivity is driven by mix shifts into lower-margin custom silicon.

The market’s reaction also signals a positioning problem: AI beneficiaries with clean supply-demand visibility are being used as liquidity sources whenever any adjacent AI headline disappoints. That creates a short window where fundamentals and flows diverge, and the best entries tend to come after forced de-grossing rather than on the initial break. The risk is that this trade becomes crowded; once investors fully re-anchor on HBM scarcity, the upside from multiple expansion fades and returns become more dependent on realized contract pricing through 2027.

The contrarian angle is that the market may be overestimating how mechanically positive inference growth is for every AI supply-chain name. If hyperscalers re-optimize compute toward lower-HBM architectures, or if next-gen packaging/stacking constraints slow unit growth, the demand curve could flatten faster than consensus expects. Still, near term the setup favors the names with sold-out capacity and contractual visibility, while the names with guidance-driven margin compression remain vulnerable to being sold on any AI mix scare.

Key Decisions for Investors

• Maintain/add to long MU on pullbacks over the next 1-3 weeks; use a stop below the prior breakout base, targeting a retest/extension above the 52-week high over 1-2 months. Risk/reward is favorable because downside is capped by sold-out 2026 capacity while upside is driven by continued AI memory tightness.
• Fade the reflex link between AVGO and HBM: if AVGO weakens on margin-mix concerns, avoid shorting MU on sympathy. Use AVGO weakness as a relative-value signal to rotate from accelerator mix exposure into memory exposure over the next 2-6 weeks.
• Pair trade: long MU / short AVGO for 1-3 months. Thesis: MU has cleaner supply visibility and direct HBM scarcity leverage, while AVGO is more exposed to mix-driven gross margin compression and multiple de-rating if investors focus on lower-margin custom silicon.
• Consider a bullish options structure on MU, such as a 3-6 month call spread financed by selling upside farther out, to capture continued momentum without overpaying for implied volatility after the recent run.
• For more defensive AI exposure, keep GS as a beneficiary of capex monetization only tactically; its role here is indirect and lower beta, so it is not the best vehicle for expressing the HBM thesis.