If You Had Invested $1,000 in Micron 1 Year Ago, Here's How Much You'd Have Now. Is the Stock Still a Buy?

Micron reported fiscal Q2 revenue of $23.86 billion, up 196% year over year, with non-GAAP gross margin expanding to about 75% and net income rising to $13.79 billion. Management guided fiscal Q3 revenue to roughly $33.5 billion, a single-quarter figure that exceeds the company’s full-year revenue for every year through fiscal 2024, while HBM supply for calendar 2026 is effectively sold out. The article is positive on fundamentals but cautions that the stock’s roughly 850% 12-month gain leaves elevated expectations embedded in the valuation.

Analysis

The market is correctly repricing MU as an AI infrastructure bottleneck, but the more important second-order effect is that HBM is turning memory from a commodity into a quasi-capacity-constrained specialty component. That shifts bargaining power upstream for the next 12-18 months: foundry and packaging ecosystems with exposure to advanced substrates, test, and assembly should remain tight, while downstream AI OEMs may see longer lead times and higher bill-of-materials pressure. The read-through is mildly negative for any AI hardware name that cannot pass through memory inflation quickly, and supportive for suppliers that participate in the HBM stack without bearing full cyclicality.

The main risk is not demand collapse tomorrow; it is supply response with a lag. Once new wafer capacity, yields, and competitor HBM qualifications normalize, margins can compress far faster than revenue does because MU’s current earnings power is heavily leveraged to pricing rather than only volume. That argues for a months-to-years horizon risk, not a days-to-weeks one: the stock can stay expensive while sell-side estimates rise, but the asymmetry worsens once the market starts discounting 2027 supply normalization instead of 2026 shortages.

Consensus likely underestimates how much of the current rally is a multiple expansion story masquerading as pure growth. A low-40s P/E on peak-ish margins is acceptable only if investors believe the HBM scarcity regime persists; any evidence of incremental supply, customer concentration, or inventory normalization would hit both the multiple and the earnings line simultaneously. On the other hand, dividend growth signals management confidence and can attract a lower-volatility ownership base, which may cushion pullbacks even if the fundamental setup remains cyclical.

The best trade is not chasing MU outright after an 8x move, but expressing the AI memory thesis with a relative-value lens. The current setup favors long supply-chain winners with less direct memory beta versus long MU, or using options to define risk while staying exposed to continued estimate revisions. If the thesis extends, upside likely comes from earnings revisions over the next 2-3 quarters; if it breaks, the unwind can be abrupt because positioning is probably crowded and sentiment is already extreme.

Key Decisions for Investors

• Avoid initiating fresh outright long MU at current levels; if already long, trim into strength and keep only a tactical core position into the next earnings cycle. Risk/reward deteriorates materially once valuation starts discounting peak margins plus 2027 scarcity.
• Pair trade: long a diversified AI infrastructure supplier basket (e.g., AMAT/LRCX/ASML depending mandate) versus short MU for 3-6 months. Thesis: the market is paying MU for scarcity, but the supply-chain toll collectors have less downside if HBM pricing normalizes.
• If needing exposure, use MU call spreads 3-6 months out rather than stock. This preserves participation in estimate revisions while capping downside if guidance already reflects near-term sellout conditions.
• Watch for a tactical short setup on MU if management commentary turns from 'sold out' to 'capacity ramp' or 'customer qualification broadening' over the next 1-2 quarters. That would be an early signal that supply is catching demand before margins roll over.
• For AI beta exposure, prefer names where memory is a cost input rather than the profit engine; they benefit if AI capex stays strong but memory pricing eases, giving better asymmetry than chasing the hottest single beneficiary.