Micron Q3 Preview: Am I Buying More? No; Am I Done?

Micron is expected to deliver 44% sequential revenue growth and 270% year-over-year growth in Q3, with the company reportedly still beating management and analyst estimates. The stock has already rallied 243% year to date, yet it trades at just 16.4x forward P/E against projected EPS of $60 in FY26 and $113.81 in FY27. The article frames AI supercycle demand and earnings momentum as the main drivers of further upside.

Analysis

MU is becoming the cleanest public-market expression of AI capex because it monetizes the memory content per server, not just unit growth. The second-order winner set extends to HBM-adjacent packaging, advanced substrates, and equipment vendors, while the losers are the PC/smartphone-sensitive memory peers that lack AI mix and may get forced into rational pricing behavior sooner. The key signal is not just demand strength but the lengthening of the cycle: if hyperscaler memory budgets stay elevated, downstream OEMs will have to de-stock less aggressively, which can keep spot/contract spreads supportive for several quarters.

The market is likely still underestimating how much earnings leverage sits in mix and utilization rather than simple bit shipments. At this stage of the cycle, even modest gross margin upside can produce outsized EPS revisions, which is why the forward multiple can look optically cheap even after a large run. The risk is that the narrative is now crowded: any sign of customer inventory normalization, capex pauses, or a temporary lull in AI server deployments can compress the multiple faster than estimates fall, especially over a 1-3 month horizon around the next print.

The contrarian view is that the consensus may be extrapolating the slope of the current AI cycle instead of the duration. Memory is notorious for peak-margin behavior, and the market often prices in several quarters of perfection right before procurement cadence slows. If the company is already being rewarded for beat-and-raise behavior, the next leg higher likely requires not just another beat, but evidence that FY27-level earnings are becoming feasible without a demand air pocket in between.

Key Decisions for Investors

• Maintain a tactical long MU into the next earnings window, but express it with a defined-risk call spread (3-6 months out) rather than common stock; target a 2:1 to 3:1 upside/downside profile if the print confirms sustained AI mix improvement.
• Pair long MU vs short a slower-moving memory or handset-exposed semi over the next 1-2 quarters to isolate AI memory demand from cyclical consumer electronics exposure; this reduces beta while keeping the idiosyncratic upside.
• If already long MU, trim 20-30% on any pre-earnings momentum spike and re-add only after management commentary on capex discipline and customer inventory levels; the main risk is multiple compression after a perfect setup.
• For event-driven accounts, buy post-earnings confirmation on a breakout above the prior high rather than chasing ahead of the report; the trade works best if guidance supports 2-3 quarters of revision momentum.
• Avoid shorting MU outright; if bearish, express it via a relative-value short against the most crowded AI beneficiary basket, because the fundamental upside skew is still positive and the strongest risk is not collapse, but time decay after a rich run.