Micron Technology — 8.4/10 — $366.24

Micron Technology is one of only three companies in the world that manufacture DRAM, the most critical memory chip in computing. Samsung (~33%), SK Hynix (~33%), and Micron (~26%) collectively control approximately 92% of the global DRAM market -- a textbook oligopoly with barriers to entry exceeding $20B for a new entrant. Micron is the only US-based memory manufacturer, giving it strategic importance for CHIPS Act funding and domestic supply chain security.

The company is in the most powerful upcycle in memory industry history. FQ2 2026 revenue of $23.9B nearly tripled year-over-year. Gross margins expanded from ~38% a year ago to 75%, with guidance for 81% in FQ3 -- a level never before achieved in memory. The FQ3 revenue guide of $33.5B exceeds the entire FY2024 annual revenue of $25.1B in a single quarter. Adjusted free cash flow of $6.9B in the first half of FY2026 already exceeds any full-year total in company history.

HBM and AI are the driving forces. High Bandwidth Memory is the critical component enabling AI accelerators (NVIDIA, AMD, custom ASICs). Micron has HBM4 in volume production -- an industry first -- with yields ramping faster than HBM3E. The company has completed agreements on price and volume for its entire CY2026 HBM supply, with some customers receiving only 50-65% of their memory demand due to supply constraints. DRAM now represents 79% of revenue (up from 70% in FY2024), and data center segments (CMBU + CDBU) reached $13.4B in FQ2 alone -- 56% of revenue. The introduction of 5-year Strategic Customer Agreements (SCAs) represents a potential structural shift from the historically transactional memory business.

The central investment question is whether this cycle is structurally different. Memory has never sustained peak margins for more than 4-6 quarters historically. From FY2022 peak ($8.35 EPS) to FY2023 trough (-$4.45 EPS) was a 152% swing. The bull case rests on three structural arguments: greenfield fab lead times of 3+ years limit supply response, HBM has a 3:1+ trade ratio that permanently constrains DRAM bit supply per wafer, and AI demand appears secular rather than cyclical. At ~10.5x FY2026E consensus EPS of ~$35, the stock is not expensive on current earnings -- but a reversion to mid-cycle earnings would change that calculus dramatically.

MU receives a composite score of 8.4/10, reflecting a company experiencing the most powerful earnings inflection in semiconductor history, with best-in-class AI thematic positioning, proven management, and favorable sentiment divergence -- tempered by the inherent cyclicality risk that has defined the memory industry for four decades.

Bull case: Micron is trading at ~10.5x FY2026E consensus EPS of ~$35, with FQ3 guidance implying an $88B annualized revenue run-rate. Revenue nearly tripled YoY in FQ2, gross margins hit 75% (guiding to 81%), and the company generated record free cash flow of $6.9B in a single quarter. HBM4 is in volume production with yields exceeding HBM3E, the entire CY2026 HBM supply is sold out, and some customers are receiving only 50-65% of their memory demand. The introduction of 5-year Strategic Customer Agreements represents a potential structural de-risking of the boom-bust memory model. Consensus price target of ~$530 implies ~45% upside. Three structural factors support sustainability: greenfield fab lead times of 3+ years limit supply response, HBM has a 3:1+ trade ratio constraining DRAM bit supply per wafer, and AI demand appears secular rather than cyclical.

Bear case: Memory has never sustained peak margins for more than 4-6 quarters. From FY2022 peak ($8.35 EPS) to FY2023 trough (-$4.45 EPS) was a 152% swing, and at prior peaks (FY2018, FY2022), MU traded at 5-8x peak P/E before correcting 50%+. Capex is guided above $25B for FY2026 with a meaningful step-up in FY2027 (construction spend up $10B+ YoY) -- history shows memory companies over-invest at cycle peaks. Samsung is investing heavily to regain HBM share. China + Hong Kong represent ~10% of revenue with export restrictions limiting advanced product sales. P/B of ~6x is well above the historical average of 2-3x. Tariff and trade policy impacts are explicitly excluded from guidance.

The key question: is this cycle different? The score of 8.4 reflects conviction that structural factors (HBM trade ratios, fab lead times, AI demand durability, SCAs) genuinely improve the risk profile compared to prior cycles -- but the score is held below 9 by the 40-year history of cyclical reversals, the aggressive capex trajectory, and the lack of transparency on SCA downside protection terms.

Key catalysts and monitoring points:

• FQ3 2026 earnings (June 2026): Guide for $33.5B revenue and $19.15 EPS. Another massive beat could drive re-rating; any miss or margin compression would raise cycle-peak concerns. This single quarter would exceed all of FY2024 revenue.
• HBM4 yield trajectory and HBM4E development: HBM4 yields are ramping faster than HBM3E. HBM4E (CY2027) features a customizable base logic die offering differentiation and higher margins. Track execution vs. SK Hynix (~62% HBM share).
• SCA expansion and terms: The first 5-year Strategic Customer Agreement was announced in FQ2. More agreements with specific commitment terms could structurally de-risk the business model. Watch for any disclosed downside margin protections.
• Capex trajectory: FY2026 guided above $25B with FY2027 stepping up meaningfully (construction spend up $10B+ YoY). This is the biggest risk factor -- if demand slows, this capex becomes a massive FCF drag. Track whether FCF growth outpaces capex growth.
• Supply-demand tightness duration: Management says tightness persists beyond calendar 2026; the Street models normalization in CY2027. Any signal from Samsung or SK Hynix on capacity additions or pricing softness is critical.
• NAND inflection: Data center SSD revenue more than doubled sequentially in FQ2. If NAND supply tightness persists, it becomes a meaningful incremental profit driver. Historically, NAND cycles are shorter and sharper than DRAM.
• Samsung HBM qualification: Samsung is investing heavily to regain HBM share (qualified at major customers for HBM3E). A Samsung comeback could pressure pricing.
• Tariff and trade policy: Management explicitly excludes tariff impacts from guidance. Escalation of US-China tensions or broader tariff actions represent unmodeled downside risk.

For the full analysis, see the Business Model, Financials, and Valuation pages.

Concerns, Catalysts & Risks -- full analysis

Buy at current levels -- Micron is a textbook oligopoly experiencing the strongest upcycle in memory history, with structural AI tailwinds that may genuinely extend the cycle beyond historical norms. At $366.24 (22% below the 52-week high of $471.34), the stock trades at ~10.5x FY2026E consensus EPS -- cheap for semis, though investors must weigh this against the reality that peak-cycle P/E multiples are inherently misleading for cyclicals.

The fundamental setup is extraordinary. Revenue nearly tripled YoY, gross margins hit levels never before seen in memory (75%, guiding 81%), HBM4 is in volume production as an industry first, and the company has sold out its entire CY2026 HBM supply with customers receiving only 50-65% of demand. Adjusted FCF of $6.9B in H1 FY2026 already exceeds any full-year total in company history. The introduction of 5-year Strategic Customer Agreements represents a potential structural improvement to the historically volatile memory business model.

Risk management: The factors that temper the BUY rating from a higher-conviction call are: (1) memory cyclicality -- this industry has never sustained peak margins beyond 4-6 quarters, and prior peaks were followed by 50%+ corrections, (2) capex above $25B in FY2026 with a meaningful step-up in FY2027 that could become a drag if demand slows, (3) Samsung investing heavily to regain HBM share, and (4) tariff and trade policy risks explicitly excluded from guidance. For investors with a 2-3 year horizon who believe in the structural thesis (HBM trade ratios, fab lead times, secular AI demand), MU presents compelling risk/reward from current levels. A pullback to $280-320 on broader market weakness or cycle-peak fears would represent an even more attractive entry point.