This semiconductor stock is up more than 200% this year. TD Cowen says it has more room to run

TD Cowen raised Micron’s price target to $1,500 from $660, implying 53% upside, as DRAM demand continues to outpace supply amid AI-driven growth. The analyst said memory’s role in AI is structural rather than cyclical and highlighted HBM as a key supply constraint due to its high capital intensity. Micron shares are up 244% year to date, and 44 of 47 covering analysts rate it buy or strong buy.

Analysis

This is less a simple cyclical upturn than a capacity-rationing regime, which means the usual “peak memory” playbook is likely the wrong framework. When a component becomes strategically embedded in AI infrastructure, pricing power can persist longer than utilization data suggests because customers optimize for supply assurance, not just spot cost. That favors the highest-quality, best-capitalized memory vendors and pushes weaker rivals into a tougher tradeoff: defend share with capex or protect margins and lose sockets.

The second-order effect is on the rest of the AI hardware stack. If memory remains the tightest constraint, hyperscalers may delay some server deployments even as they continue ordering accelerators, creating intermittent demand spikes for memory while making compute suppliers more sensitive to build timing. It also supports a broader read-through to suppliers of advanced packaging and equipment: the more expensive and capital-intensive the bottleneck becomes, the more the industry shifts toward oligopoly economics and away from classic boom-bust semis behavior.

The main risk is that the market is already discounting a multi-quarter scarcity narrative, so any soft patch in enterprise IT or a pause in hyperscaler digestion could trigger an air pocket even without a true fundamental rollover. The reversal signal is not just price weakness, but evidence that customers are substituting mix, carrying excess inventory, or rephasing AI capex into next year. On the contrarian side, the consensus may be underestimating how much of the upside has already been captured in the stock after a very large year-to-date move; the cleaner edge is likely in relative value and defined-risk expressions rather than outright chase buying.

Key Decisions for Investors

• Stay long MU, but express it via call spreads rather than stock for the next 3-6 months; target another leg higher if AI capex re-accelerates, while capping downside if the trade becomes crowded.
• Pair trade: long MU / short a weaker memory or capital-intensive semi supplier over 1-2 quarters; the thesis is that scarcity economics accrue to the most disciplined operator while laggards face margin compression.
• Use a pullback of 5-8% in MU to add, not chase strength; the stock is in a momentum regime, but after a large rerating the best entry is likely on inventory/digestion-related weakness.
• For hedging, buy downside puts on MU into any earnings or guidance event over the next 1-2 quarters; the key risk is not demand collapse but a sentiment reset if the market prices in an earlier-than-expected peak.
• Watch server OEM and hyperscaler commentary over the next 30-60 days; any mention of memory allocation tightness or lead-time extension is a positive confirmatory signal for keeping the long on.