TD Cowen raises Micron stock price target on margin outlook

TD Cowen raised Micron’s price target to $660 from $550 and maintained a Buy rating, citing AI-driven demand and long-term margin durability. The firm sees EPS of $23 in May and $27 in August versus $19 and $23 consensus, while also highlighting potential near-term headwinds as margins normalize from high levels. Micron is separately facing strategic attention around U.S. export controls, but the core takeaway is continued bullish analyst sentiment on AI memory demand.

Analysis

The key incremental signal is not the near-term beat potential; it is that the market is beginning to anchor MU’s terminal margin structure well above historical troughs even while the stock is already discounting a reset. That creates a classic “two-part tape”: upside can continue on earnings revisions, but the more important question is whether duration gets extended enough to justify multiple expansion into the next cycle. If AI memory demand stays tight, semis tied to training spend can tolerate a higher valuation regime because memory supply discipline tends to lag demand by multiple quarters.

Second-order beneficiaries are the equipment and materials vendors that monetize capacity additions without needing perfect end-market visibility. If Micron’s margin floor is truly migrating upward, the next leg of capex should flow through to lithography, deposition, test, and packaging names before it shows up in consensus revenue, creating a cleaner risk/reward than chasing MU after a large move. Conversely, any sign that DRAM pricing is peaking earlier than expected would hit the whole AI memory complex at once because the current setup is built on margin convexity, not just volume growth.

The main contrarian point is that consensus may be underestimating how quickly the market can re-rate a memory name when investors stop treating it like a cyclical and start treating it like a scarcity asset. That said, the stock is vulnerable to a sharp de-rating if gross margins merely revert rather than re-accelerate, because the multiple is already implying confidence in a benign post-2027 curve. The risk window is the next 1-2 earnings prints: any guidance wobble would matter more than the reported quarter because the narrative is carrying much of the valuation.

NVDA is a secondary loser if the market interprets stronger memory economics as more evidence that GPU supply-chain bottlenecks are easing rather than broadening the AI pie. Export-control rhetoric adds a separate layer of policy optionality: tighter restrictions could support domestic memory and equipment names while compressing China-exposed semiconductor duration elsewhere. In short, the trade is less about owning MU outright and more about owning the supply-chain beneficiaries while using MU as a sentiment barometer for the AI buildout.

Key Decisions for Investors

• Stay long MU into the next print, but monetize strength with a defined-risk collar: own the equity and buy 1-2 month downside protection 5-8% below spot; the setup favors upside if revisions continue, but the multiple can de-rate fast on any margin-compression warning.
• Pair trade: long MU / short NVDA for 1-2 quarters if the market starts pricing AI memory scarcity as the cleaner beneficiary of incremental capex; the relative trade works if memory upside is driven by pricing power rather than broad AI demand acceleration.
• Buy semicap exposure on pullbacks over the next 2-6 weeks: use AMAT/LRCX as higher-quality second-order beneficiaries of capacity expansion; risk/reward is better than chasing MU after a strong run because capex tends to follow pricing with a lag.
• If MU trades above the new target area on no new data, take profits on 25-33% of any long and wait for the next guide cycle; the stock is already discounting a favorable long-run margin structure, so incremental upside should be funded by earnings revisions, not multiple expansion.
• For a tactical catalyst trade, consider MU call spreads 1-2 months out into earnings; positive skew if estimates are again beaten, but cap premium paid because the main downside is not earnings miss alone, it is any signal that the gross margin floor is softer than the market assumes.