Micron Crushes Earnings: Is a Stock Split Next?

Micron posted a dramatic fiscal Q3 surge, with revenue rising from $9.3B to $41.5B and gross margin expanding from 37.7% to 84.6%, driving adjusted EPS to $25.11 from $1.91. The move is attributed to AI-driven DRAM/NAND supply-demand imbalances (including HBM for GPUs and flash/SSD demand for AI data centers) and improving pricing power, with ~40% of revenue tied to new long-term contracts. A stock split is floated as a possible near-term catalyst, but the article highlights further contract tie-ups as the more important lever to reduce cyclicality; shares have still lost ~25% after an earlier surge to an all-time high.

Analysis

The market is mispricing the catalyst stack by focusing on a cosmetic split and underweighting the real issue: whether Micron can convert spot-driven HBM/NAND scarcity into durable contract pricing. If long-term agreements keep rising from here, MU’s multiple can re-rate from a cyclical 6x forward earnings toward a semi-structural AI supplier multiple; if not, the current margin profile is the kind that usually attracts supply within 2-3 quarters.

Second-order winners are the memory equipment and packaging ecosystem (AMAT, LRCX, KLAC, likely OSAT/advanced packaging chains) because any attempt to lock in supply implies more capex and process complexity. The underappreciated loser is the broader AI hardware stack: NVDA, server ODMs, and hyperscalers get the same compute economics only if HBM remains available, but sustained memory inflation can squeeze system-level margins before it shows up in headline AI demand.

Contrarian view: the stock may already be telling us that earnings are peak-ish, not because the AI thesis is broken, but because the street is discounting eventual capacity response from SK Hynix and Samsung. The key falsifier is not stock-split chatter; it is whether contract revenue can move materially above 40% and whether gross margin holds into the next pricing reset. If backlog doesn’t extend and DRAM spot rolls over, MU can de-rate quickly despite still-strong reported EPS.

Time horizon matters: near term, any split announcement is a retail-liquidity event; 1-3 months, the real catalyst is guidance on mix/contracting; 6-18 months, the thesis lives or dies on whether AI memory becomes recurring rather than cyclical. This is one of the few large-cap semis where the balance between scarcity and supply response is more important than the absolute earnings print.

Key Decisions for Investors

• Buy MU on 5-10% pullbacks, but only if management indicates higher contract mix or commentary on HBM allocation stays tight; upside is multiple expansion, not another one-quarter EPS beat.
• Pair trade: long MU / short WDC for a 3-6 month window; MU has better HBM exposure and balance-sheet quality, while WDC is more exposed to NAND normalization and lower pricing power.
• Watch for a stock-split announcement as a trading catalyst only; if MU rallies >8-10% on split news without a guidance revision, fade part of the move because the signal is liquidity-driven, not fundamental.
• Set a thesis-falsifier alert on MU gross margin and contract revenue at next print: if contract mix stalls below ~40% or margins roll over, cut the long immediately.
• For AI hardware exposure, prefer AMAT/LRCX over beta-heavy semis if the goal is to own the capex wave rather than the memory price cycle.