Why Micron Stock Bounced Back Today

Wolfe Research raised its price target on Micron to $1,250 per share, implying about 26% upside from current levels, and sees the bull case extending for another couple of years. Analyst Chris Caso expects DRAM prices to be 200% higher and NAND prices 216% higher by the end of 2026 versus end-2025, supported by insatiable AI data center demand. Wolfe also sees Micron potentially generating $226.5 billion in sales and $135 in EPS next year, with pricing strength possibly lasting into 2028 or 2029.

Analysis

The important read-through is not just that memory prices are firm; it is that the AI capex cycle is now pulling DRAM/NAND into a quasi-utility pricing regime where supply discipline matters more than end-demand headline noise. That favors the highest-cost, least-flexible producers and punishes any peer that needs aggressive wafer starts or capex to defend share, because incremental supply would immediately compress pricing power. The second-order winner is the broader AI server stack: if memory remains tight, OEMs and cloud buyers will be forced to optimize configs, which extends replacement cycles for some non-memory components while keeping spending concentrated in the highest-return compute deployments.

The market is likely underestimating the duration risk on the upside and the tail risk on the downside. Upside can persist for years if producers keep supply rational, but that discipline is fragile; the first sign of a capacity race would be a sharp re-rating lower well before spot prices roll over. The key reversal catalyst is not weak AI demand, but a change in industry behavior: one or two large players deciding to capture share at the expense of margins would break the thesis faster than any macro slowdown.

The consensus is probably too linear on earnings power. If pricing extends into 2027-2028, forward estimates will likely lag reality, creating repeated upward revisions and multiple expansion; but that also means the stock becomes increasingly vulnerable to any guidance that implies peak margins are being extrapolated too far. The more interesting contrarian angle is that the beneficiaries may actually be the equipment and infrastructure vendors that monetize the capex boom without direct exposure to memory price mean reversion.

Key Decisions for Investors

• Stay long MU, but express it with a staggered entry and trailing stop: add on pullbacks over the next 2-6 weeks, keep a 15-20% stop from cost, and look for 12-18 month upside as estimate revisions catch up to pricing.
• Pair trade: long MU / short a second-tier memory peer with more balance-sheet or capex pressure. The thesis is that pricing power accrues disproportionately to the lowest-cost operator, while weaker peers will be forced to chase supply and destroy margin.
• Buy NVDA on memory-tightness dips rather than momentum highs: if HBM remains constrained, gross margin risk is more about allocation than demand, and any selloff on memory headlines is a better entry than chasing upside.