Micron Technology Inc. (MU): Morgan Stanley stays Overweight on Micron ahead of earnings as AI-driven memory tightness is expected to persist

Overview

This weekly Micron earnings preview argues that memory-market conditions remain exceptionally strong, though the expected upside at the upcoming print is more moderate than in prior quarters. Morgan Stanley maintains Overweight and its US$1,200 target, emphasizing AI demand, supply constraints and the prospect of a multi-year tight market.

Core views

Morgan Stanley expects Micron's upcoming earnings release to produce upward revisions, but of a smaller magnitude than in recent quarters. Near-term conditions remain strong: DRAM demand and pricing strength are already consensus expectations, while NAND expectations have also moved higher following sharp TrendForce revisions. The firm sees calendar fourth-quarter pricing beginning up 15-20%, with some prices rising more, although it notes incomplete visibility into the full quarter. For the August quarter, Morgan Stanley models revenue of US$50.024bn, up 20.7% quarter on quarter and 342.1% year on year, below Street expectations of US$50.923bn. Its segment assumptions are DRAM bit shipment growth of 3.0% quarter on quarter and DRAM pricing up 16.0%, plus NAND bit shipment growth of 4.0% and NAND pricing up 20.0%. The firm's 86.4% gross-margin forecast exceeds the Street's 85.3%, while its US$31.20 EPS estimate is slightly below the Street's US$31.49. For the November quarter, Morgan Stanley expects the earnings setup to improve. It forecasts revenue of US$57.560bn, up 15.1% quarter on quarter and 321.9% year on year, above consensus of US$56.644bn. Its forecast assumes DRAM bit growth of 3.0% and pricing growth of 12.0%, as well as NAND bit growth of 4.0% and pricing growth of 10.0%. Gross margin is projected at 88.1%, above the Street's 85.3%, and EPS at US$35.57, above consensus of US$35.07. The firm characterizes the projected 10-12% like-for-like pricing increase as conservative, while noting that Micron typically guides conservatively. The report's larger thesis is that the market is now debating the longevity, rather than the peak, of the memory cycle. Morgan Stanley expects multiple years of very tight supply-demand conditions and believes AI compute deployments remain constrained by DRAM availability. In its view, customers will procure available memory with relatively low price sensitivity in order to deploy compute, while adjustments that reduce one form of memory demand can create demand elsewhere. For example, reduced LPDDR5 memory in NVIDIA racks may indirectly increase NAND caching needs. Disaggregation and networking may improve memory optimization, but the firm believes additional supply would still be absorbed quickly. Morgan Stanley addresses investor concerns around de-specification, Chinese supply and rising capital spending. It considers Chinese supply a long-monitored issue but expects it will not outpace Western vendors in the current environment, particularly as shortages are also acute in China. It has raised DRAM supply forecasts as capital spending increases, while NAND supply has been steadier; nevertheless, industry contacts expect supply-demand conditions to be tighter in 2027 and 2028 than today. The firm believes the cycle is more likely to end if AI spending slows than because of incremental supply. Micron's AI and HBM execution is central to the valuation case. Morgan Stanley argues that HBM's greater wafer intensity tightens overall memory supply-demand, while Micron's ability to maintain HBM share in calendar 2026 versus competitors supports margins and a higher valuation multiple than in prior cycles. Its base case uses 30x through-cycle earnings of US$40, a premium to its US$38 average for FY2018-FY2028 and a multiple it considers the midpoint of the semiconductor group, reflecting the AI opportunity. The bull case is US$1,650, based on 33x through-cycle earnings of US$50; the bear case is US$675, based on 27x through-cycle earnings of US$25. Morgan Stanley maintains Overweight and its US$1,200 target, while noting it favors compute names over Micron. It views the durability argument as likely to take longer to be reflected because it cannot be conclusively demonstrated in the near term. The firm also tempers expectations for share repurchases: cash returns have been limited by CHIPS Act-related constraints, and although Micron should have greater buyback capacity in December, the scale of any action remains uncertain.

Analysis framework

Morgan Stanley compares its quarterly revenue, shipment, pricing, margin and EPS forecasts with consensus, then links the earnings outlook to DRAM and NAND supply-demand conditions. It tests the durability thesis against investor concerns about AI memory requirements, Chinese supply, capital spending and end-demand risk, and values Micron using through-cycle earnings scenarios and price-to-earnings multiples.

Methodology notes

• Industry AnalysisSupply-demand framework

  Memory supply-demand analysis

  The report assesses DRAM and NAND pricing, shipments, supply additions, enterprise deployment and AI-driven demand to judge the duration of the memory upcycle.

• Valuation methodsP/E and PEG Valuation

  Through-cycle price-to-earnings scenario valuation

  Morgan Stanley derives its base, bull and bear price cases by applying 30x, 33x and 27x multiples to through-cycle earnings assumptions of US$40, US$50 and US$25, respectively.

• Event-Driven and Behavioral FinanceEvent-driven analysis

  Earnings-event and options-implied scenario analysis

  The report frames the upcoming earnings release as a catalyst and references options-implied, risk-neutral probabilities for its bull, base and bear price scenarios.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Micron Technology Inc. (MU.O)

  Primary covered company; expected to benefit from strong DRAM and NAND pricing, AI demand and HBM execution.

  Strengths

  Improving pricing, AI and datacenter exposure, expected HBM-share resilience, and projected margin expansion.

  Weaknesses

  The durability of the upcycle is difficult to prove in the near term, and buyback scale remains uncertain.

  Comparison

  Morgan Stanley views its 30x base multiple as the midpoint of the semiconductor group and states that it favors compute names over Micron.

  Risks

  A slowdown in AI spending, an end-demand shortfall with elevated inventories, HBM demand weakness or intensified competition could pressure pricing and valuation.

Key data

• August-quarter revenue estimateUS$50.024bnUp 20.7% q/q and 342.1% y/y; below Street consensus of US$50.923bn.
• August-quarter gross margin estimate86.4%Above Street consensus of 85.3%.
• August-quarter EPS estimateUS$31.20Below Street consensus of US$31.49.
• November-quarter revenue estimateUS$57.560bnUp 15.1% q/q and 321.9% y/y; above Street consensus of US$56.644bn.
• November-quarter gross margin estimate88.1%Above Street consensus of 85.3%.
• November-quarter EPS estimateUS$35.57Above Street consensus of US$35.07.
• November-quarter pricing forecast10-12%Expected like-for-like pricing increase.
• Base-case valuationUS$1,200Based on 30x through-cycle earnings of US$40.
• Bull and bear casesUS$1,650 / US$675Bull case uses 33x US$50 through-cycle earnings; bear case uses 27x US$25.

Impact & implications

The report argues that AI-led demand, HBM wafer intensity and constrained supply should support Micron's pricing, margins and earnings beyond the near-term print. It expects the market's focus to increasingly rest on proof that the cycle can remain tight for two to three more years, rather than on the size of the next quarterly beat.

Risks

• Long-term agreements cover probably less than 30% of bits and contain price caps, potentially limiting the immediate benefit of higher spot and contract pricing.
• A 14-week August quarter could create modest headwinds.
• Memory pricing could decline quickly if end demand weakens while customer inventories are elevated.
• HBM demand could falter and competition could intensify, pressuring pricing.
• The bear case assumes memory enters a downturn in early 2027 because apparent demand proves to have been customer inventory building.
• The cycle could end if AI spending slows.

What to watch

• Micron's September 30 earnings report and its November-quarter guidance.
• DRAM and NAND pricing, including whether calendar fourth-quarter increases meet the expected 10-12% range.
• Evidence that enterprise memory purchases continue to be deployed immediately.
• Micron's HBM share and competitive execution in calendar 2026.
• The scale of potential share repurchases once Micron has greater flexibility in December.
• Chinese memory supply, capital-spending-driven supply additions and signs of inventory accumulation.