Micron: HBM scarcity pricing meets clean-room constraints through 2028, not a cycle top

Stevie AI on Micron Technology, Inc. (MU-USA | microntechno)

10/27/2026

Summary

Micron manufactures DRAM and NAND memory serving data center, client, mobile, and embedded markets, and is now riding the most favorable supply-demand setup in its history: HBM's rapid mix shift, rising HBM3-to-HBM4 trade ratios, and industry-wide clean-room scarcity have converted memory from a commodity cycle into a multi-year structurally tight market. The key insight the market is underpricing is that this is not a typical memory spike destined for immediate reversal; management has already booked over 75 percent of FY2027 shipments under long-dated supply agreements, and the primary bottleneck (clean-room construction, with new capacity not online until late 2028) is a multi-year physical constraint rather than a pricing decision any single competitor can quickly undo. The historical numbers confirm the magnitude of the inflection: revenue grew from $37.4B in FY2025 to $133.2B in FY2026, net income exploded from $8.5B to $85.0B, and EPS rose from $7.59 to $74.33, with gross margins reaching 80.7 percent and net margins of 63.8 percent in FY2026. This is a step-change in profitability driven by HBM mix and tight DRAM/NAND pricing, not merely a seasonal upswing, and management guides to further margin expansion in FY2027 after an initial cost-inflation quarter. Our price targets are built on EPS of $98.87 (FY2027), $116.52 (FY2028), $117.58 (FY2029), and $112.33 (FY2030), applying a 12x forward P/E that reflects a cyclical semiconductor business with extraordinary but ultimately moderating margins, net cash building toward $278B by FY2030, and a bit-growth algorithm still structurally supported by AI server demand. At 12x FY2027 EPS, our target of approximately $1,186 implies meaningful upside from the current $903.7 price, with the path through FY2028 offering the highest absolute earnings power before gradual normalization sets in.

Thesis

1. Structural Scarcity, Not Cyclical Spike: Micron's current profitability is unusual among memory upcycles because the supply constraint is clean-room space, a multi-year construction bottleneck, rather than excess equipment capacity waiting to be switched on. Management has been explicit that new clean rooms will not come online until late calendar 2028, meaning the industry cannot meaningfully respond to high prices with new supply for at least two more years. This timeline gives Micron unusually high revenue and margin visibility relative to prior cycles, evidenced by management's disclosure that over 75 percent of FY2027 shipments are already committed under supply agreements. 2. HBM Mix Shift as a Structural Margin Driver: As the industry transitions from HBM3 to HBM4/HBM4E, trade ratios (the amount of conventional DRAM capacity consumed to produce a given amount of HBM output) are rising, which mechanically tightens conventional DRAM supply even without demand growth. Micron's HBM market share has reached parity with its overall DRAM share, positioning it to capture this mix shift proportionally rather than being squeezed out by Samsung or SK Hynix. This dynamic supports our view that pricing power persists well beyond a single product cycle. 3. Financial Trajectory Supports Premium Current Earnings Power: The forecast path shows revenue growing from $133.2B (FY2026) to a peak near $215.6B (FY2030), with FCF generation of $70-81B annually in the FY2027-2029 window and net cash accumulating from $97.9B to $278.2B. This balance sheet transformation, from a capital-intensive cyclical business to a self-funding cash generator, sets up capital returns (buybacks) once net cash is firmly established, which management has signaled as a priority once the cash position allows it. 4. Market Is Mispricing Durability, Not Magnitude: Investors are familiar with memory cycles and tend to apply trough multiples to peak earnings, assuming imminent mean reversion. However, the specific combination of AI-driven server demand growing in the high teens, diminishing bit-growth returns from node transitions, and the multi-year clean-room bottleneck suggests this upcycle has a longer duration than consensus assumes. Our 12x forward multiple reflects this: it is well below growth-stock multiples but above trough-cycle multiples of 5-7x, capturing the extended but not permanent nature of current margins. 5. Competitive Position Is Defensible Near-Term: Micron, Samsung, and SK Hynix collectively control approximately 90 percent of the DRAM market, and Micron maintains roughly a two-node technology lead over Chinese competitors. This oligopolistic structure, combined with long-dated customer lock-in via strategic commercial agreements (SCAs), gives Micron pricing visibility that is unusual for a historically commodity business. 6. Catalysts Reinforce the Thesis Over the Next 12 Months: Q1 FY2027 earnings in late November 2026 will test whether single-digit sequential bit growth guidance holds and margin expansion continues as promised. The Idaho fab's first wafer output in mid-2027 will be an early proof point of successful domestic capacity execution, and continued clean-room construction updates will inform the market's view of when (if ever before FY2029) supply normalization begins.

Risks

1. Cyclical Pricing Reversal: Micron's extraordinary 63.8 percent net margin in FY2026 depends entirely on the current supply-demand imbalance. Any faster-than-expected resolution of clean-room constraints, acceleration of competitor capacity additions, or softening of AI capex could compress pricing and margins rapidly, as has happened in every prior memory cycle. 2. Execution Risk on Capacity Timing: The thesis depends on clean-room construction timelines holding through late 2028. Construction delays, equipment supply chain issues, or faster-than-guided ramp by Samsung or SK Hynix could alter the supply timeline in either direction, with faster competitor ramps being the more negative scenario for Micron's pricing power. 3. Customer Concentration and SCA Renegotiation Risk: While long-dated supply agreements provide revenue visibility, they also mean Micron's largest customers (hyperscalers and AI infrastructure buyers) have significant negotiating leverage at renewal. A pullback in AI capex spending by a small number of large customers could disproportionately affect Micron's bit demand assumptions embedded in the high-teens server growth forecast. 4. Margin Normalization Already Embedded Starting FY2029: Our own forecast shows net income declining from $136.2B (FY2029) to $129.6B (FY2030) and EPS falling from $117.58 to $112.33, reflecting management's own acknowledgment that price increases will eventually moderate. If this normalization happens faster or more severely than modeled, FY2029-2030 estimates and the corresponding price targets would need to be revised downward. 5. Capital Intensity and Execution on Capex: The company is running an elevated, multi-year capex program tied to clean-room construction and equipping. Any missteps in capital allocation, cost overruns, or misjudging demand at the time equipment decisions are made (per management's stated approach of equipping in line with demand trends observed at the time) could strand capital or delay the capacity response needed to sustain growth into the 2030s. 6. China Competitive Catch-Up: Although Micron maintains an approximate two-node technology lead over Chinese memory competitors today, continued state-backed investment in Chinese DRAM and NAND capacity represents a multi-year risk to pricing and market share, particularly in conventional (non-HBM) memory segments where technology differentiation is less pronounced.

📈 Price Targets