One AI Stock I'd Watch Before Nvidia's Earnings

Micron (MU) is framed as the lead HBM supplier in the AI memory trade, with a 12-month price target of $954.59 vs. $910.43 (4.85% upside) and a HOLD rating at 90% confidence. The article cites strong fundamentals—fiscal Q3 revenue of $41.5B (+346% YoY) and non-GAAP EPS of $25.11 (+23.79% vs. consensus)—plus aggressive Q4 guidance of $50B revenue and $31 EPS (midpoint). Key bull catalysts are tight DRAM/NAND dynamics and ramping HBM4 (>$1B revenue booked), but risks center on rapidly rising capex (Q3 spend $7.826B; ~ $10B guided for fiscal Q4) and Nvidia/AI capex sensitivity due to HBM4 concentration.

Analysis

MU is no longer a clean “AI scarcity” trade; it has become a consensus proxy for the entire AI memory bottleneck, which means the stock now trades as much on confidence in 2026-27 supply discipline as on near-term orders. The key mechanism is margin durability: if HBM pricing stays tight, MU’s high fixed-cost base and contracted mix can keep earnings leverage extreme; if pricing cracks even modestly, the multiple can de-rate much faster than fundamentals because the market is already discounting a quasi-annuity stream.

The main second-order risk sits upstream and downstream. Upstream, the current capex surge raises the odds that industry supply catches demand later than bulls expect, which would pressure DRAM pricing first and HBM second; downstream, any softening in NVDA’s tone would not just hit MU sentiment, it would also impair the willingness of customers to sign long-duration take-or-pay structures at premium margins. WDC is a relative beneficiary only if investors rotate toward cheaper AI-infrastructure exposure; it is not a true hedge if the AI buildout narrative breaks.

The contrarian point is that the stock may already be pricing the best imaginable version of the cycle while still leaving execution risk underappreciated. The near-term catalyst is NVDA guidance this week; the 1-3 month path is whether MU can show that HBM4 ramp and booked backlog convert into pricing power rather than merely volume growth. Over 6-18 months, the falsifier is simple: any evidence of HBM4 ASP softness, capex moderation, or a smaller-than-expected AI server budget at hyperscalers would likely compress MU’s multiple well before earnings estimates roll over.

Key Decisions for Investors

• Buy a small NVDA/MU pair: long NVDA, short MU into the NVDA print, targeting relative underperformance in MU if guidance merely inlines; thesis invalidates if MU re-accelerates after earnings on explicit HBM4 pricing commentary.
• For event risk, consider a 1-2 week MU put spread rather than outright short equity; the setup is asymmetric because the market is already paying for perfection and post-earnings moves in mature consensus names can gap lower on any guide ambiguity.
• If you want AI infrastructure exposure without the full HBM crowding, rotate part of MU exposure into WDC on a relative-value basis only after confirming AI storage demand remains intact; this is a lower-conviction expression and should be treated as a valuation trade, not a macro hedge.
• Set an alert on NVDA capex commentary and MU HBM4 pricing signals; if NVDA implies slower-than-feared 2026 demand, de-risk MU immediately as the multiple compression risk will likely outrun estimate cuts.
• Use $1,000 as a tactical resistance level in MU: upside through that area would suggest the market is still underestimating scarcity, but failure to reclaim it after NVDA would argue for taking profits on long exposure.