Micron Technology
The AI-Memory Super-Cycle

Snapshot: Executive Summary

Micron just printed the strongest quarter in its history. Q3 FY26 (reported Jun 24, 2026) delivered $41.46B revenue (+346% YoY, +74% QoQ)¹, a record non-GAAP gross margin of 84.9%, and non-GAAP EPS of $25.11: an all-time record that beat the ~$35.8B revenue consensus by ~$5.6B. Management guided Q4 FY26 to a record $50.0B ± $1.0B² with ~86% GM and $31.00 ± $1.00 EPS.

Drivers:

• An HBM/AI-memory super-cycle: data-center revenue surpassed $25B (a >$100B annualized run-rate), with HBM + high-capacity DIMMs + LP server DRAM together ~$10B (5x+ YoY).
• HBM4 already in high-volume shipments, >$1B revenue, with the 12-high ramp tracking ~2x faster than HBM3E; HBM share ~21%¹⁰ (Q1 2026, tied with Samsung behind SK Hynix), qualified on NVIDIA and AMD¹².
• 16 Strategic Customer Agreements ≈ $100B of multi-year take-or-pay revenue now cover roughly half of total revenue, a structural de-risking absent in prior cycles.

Net cash hit a record +$24.4B⁵ (cash + investments $30.1B less total debt $5.72B, after a $4.4B debt paydown); record FCF was $18.3B. FY26 capex is ~$27B, with FY27 stepping up to the mid-$40Bs², funded entirely from the balance sheet. At ~11.5x FY27E earnings ($98.52, +172% YoY) vs SOX (~28x), MU remains the cheapest name in the AI-memory complex on growth despite a +259.2% YTD rally. Reiterate Buy, 12-mo PT $1,275 (~+12.6% upside; in line with the ~$1,264 consensus mean; bull $1,680 / bear $620).

Tactical: MU is trading at $1,132, ~11% below our $1,275 12-month target after a -6.69% session off the Jun 25 52-week high. At only ~11.5x FY27E EPS, the gap reflects classic memory-cycle skepticism more than fundamentals: ~$100B of take-or-pay SCAs now cover ~half of revenue and net cash is a record +$24.4B. Further upside toward the $1,680 bull case requires Micron to hold flagship NVIDIA Vera Rubin HBM4 allocation and FY27 EPS to move toward ~$115; the $620 bear case reflects a SK-Hynix-led HBM4 lock-out and a conventional-DRAM reset. Rating Buy: but with the stock at peak-cycle EPS and +259% YTD, size with discipline around the blended ~$1,266 fair value.

Investment Thesis Summary

Overall Rating: Buy · 12-month PT $1,275

We reiterate Buy on Micron with a 12-month price target of $1,275 (12.9x FY27E non-GAAP EPS of $98.52), ~+12.6% above the $1,132.33 spot and in line with the ~$1,264 sell-side consensus mean. The stock has rallied +259.2% YTD; at ~11.5x forward earnings MU is still the cheapest name in the AI-memory complex on growth, well below SOX (~28x) and NVDA (~35x). This cycle is structurally different: HBM and contract-based AI-memory demand have compressed the historical boom-bust risk, ~$100B of multi-year take-or-pay Strategic Customer Agreements now cover roughly half of revenue, and the balance sheet carries a record +$24.4B net cash. The rating is Buy rather than Strong Buy because the stock is at peak-cycle EPS/margins and the single biggest swing factor, whether Micron holds flagship NVIDIA Vera Rubin HBM4 allocation versus being confined to inference-tier Rubin CPX, is genuinely contested (reporting skews SK-Hynix), which is why the bear case carries a 20% weight and we counsel disciplined sizing on a +259% YTD move.

1 · Business Overview

Product Segments (Q3 FY26)

Micron reports across four business units: Cloud Memory (HBM, hyperscaler DRAM), Core Data Center, Mobile & Client, and Auto/Embedded. In Q3 FY26 the DRAM/NAND split was DRAM $31.3B (76%) and NAND $9.9B (24%, ~doubled QoQ). Data-center revenue surpassed $25B (a >$100B annualized run-rate).

End Markets (Q3 FY26)

• Cloud Memory ($17.31B, 83% GM): HBM + hyperscaler DRAM. The fastest-growing and largest BU, driven by the HBM3E/HBM4 ramp.
• Core Data Center ($11.52B, 87% GM): High-capacity server DRAM + DIMMs + enterprise SSD. Highest-margin BU.
• Mobile & Client ($8.00B, 73% GM): LPDDR5X/client SSD, unit growth plus sharply higher ASPs, at a structurally lower margin than the data-center BUs.
• Auto & Embedded ($4.63B, 79% GM): Highest content-per-vehicle in the industry; ADAS/infotainment secular growth.

Value-Chain Position

Memory is a 3-player oligopoly (Samsung, SK Hynix, Micron control >95% of DRAM bit supply; Q1 2026 DRAM share Samsung 38% / SK Hynix 29% / Micron 22%). Micron is uniquely the only US-domiciled producer, with CHIPS-funded Idaho/NY fabs. In HBM it went from non-participant to a ~21% share (Q1 2026, tied with Samsung behind SK Hynix at 58%), is qualified across NVIDIA and dual-sourced on AMD MI350X/MI355X, and is now shipping HBM4 in high volume.

Competitor Deep-Dive

1b · Operating Drivers: Bits, Cost, and Footprint

Bit Growth & ASP Trajectory

The memory cost curve: Historically, DRAM cost-per-bit has fallen ~15% annually (Moore's Law for memory) while ASP/bit has been volatile. Margin expansion happens when ASP rises faster than cost, and in Q3 FY26 that gap was the widest on record, lifting non-GAAP gross margin to 84.9%. HBM4 amplifies the ratio further: ASP-per-bit runs well above conventional DRAM (HBM uses ~3x die area per Gb), so each HBM wafer cannibalizes conventional DRAM supply and tightens the whole market.

Manufacturing Footprint

~70% of leading-edge wafer starts are in Taiwan. CHIPS Act-funded Boise + NY ramps shift mix toward US over 2027-2030. Geographic diversification is a strategic priority.

Customer Concentration

Customer concentration is the highest in MU's history (typically 35-40% pre-HBM era), the natural consequence of HBM being a hyperscaler-exclusive product. The mitigant this cycle is the 16 Strategic Customer Agreements (~$100B of minimum contracted, take-or-pay, price-band revenue covering ~20% of DRAM and up to 1/3 of NAND volume), targeting half-or-more of total revenue under contract, a structural offset to concentration absent in prior cycles.

Geographic Revenue Mix (Q3 FY26)

2 · Memory Cycle Analysis

Where We Are

This is the steepest memory up-cycle on record. Q3 FY26 non-GAAP gross margin of 84.9% sits far above prior cyclical peaks (~62% in CY2018). Critically, this is not a typical commodity rally, supply is constrained by HBM cannibalization of conventional DRAM wafers (HBM uses ~3x die area per Gb), HBM3E+HBM4 are fully booked through CY2027, and ~$100B of take-or-pay SCAs underpin roughly half of revenue. Tight supply persists through 2027; meaningful new greenfield bits only arrive in 2028.

This Cycle vs Prior Cycles

Cycle by Product

2b · AI & Data Center Catalyst

HBM3E → HBM4 Positioning

Micron skipped HBM3 and went directly to HBM3E (24GB 8-Hi into volume production for NVIDIA H200 in 2024, 12-Hi for B200/B300). It is now shipping HBM4 36GB 12-high in high volume: already >$1B in revenue, with the 12-high ramp tracking ~2x faster than HBM3E, and HBM4E volume targeted for CY2027 on the 1-gamma node. In Q3 FY26, HBM + high-capacity DIMMs + LP server DRAM together were ~$10B (5x+ YoY).

HBM Product Roadmap (Calendar Years)

Hyperscaler Capex

HBM Market Share

• Installed share (Q1 2026, TrendForce/Counterpoint): SK Hynix 58% · Micron ~21% · Samsung 21%, Micron and Samsung are tied for the #2 slot behind SK Hynix.
• Micron trajectory: ~21-23% base case; the bull case lifts share toward ~25% on sustained flagship Rubin HBM4 allocation.
• Samsung: Began HBM4 shipments in Feb 2026 and is a credible second source for Vera Rubin, directly contesting Micron's #2 position.
• HBM4 status: Micron is in high-volume shipments of HBM4 36GB 12-high (>$1B revenue already; 12-high ramp ~2x faster than HBM3E); HBM4E volume targeted CY2027 on 1-gamma DRAM.
• Forward HBM4 allocation (the swing factor): Reporting on the NVIDIA Vera Rubin HBM4 split (SK Hynix ~55-60%+ / Samsung ~25-30% / Micron the remainder) suggests Micron could be confined to inference-tier Rubin CPX rather than flagship Vera Rubin. This is the single biggest downside swing and the reason the bear case carries a 20% weight; Micron's own framing (HVM, >$1B HBM4 revenue, 12-high ramp) is more bullish than third-party allocation reporting.

Customer Relationships

Google TPU socket included for thesis-graph consistency, Samsung supplies ~60%+ of Google TPU HBM through the Broadcom co-design relationship; Micron's near-term TPU qualification is constrained by wafer capacity vs Korean peers, with HBM4 qualification windows the next decision point.

3 · Financial Health & Trends

Quarterly Summary

Margin Read

• Cycle peak: Q3 FY26 non-GAAP GM (84.9%) is an all-time corporate record, far above the prior cyclical peak (~62% in CY2018); the Q4 guide of ~86% would set another. This is peak-cycle margin math, the central reason the rating is Buy rather than Strong Buy.
• Operating leverage: Non-GAAP operating margin reached 81.2% (GAAP operating income $33.32B, 80.4%). Incremental drop-through on the revenue ramp is extraordinary as fixed costs are spread over a 74% sequential revenue gain.

Balance Sheet (May 28, 2026)

3b · Capital Allocation & Returns

Capital Return Policy

Capital priorities (mgmt-stated):

• Reinvest in HBM / leading-edge DRAM (FY27 capex stepping to the mid-$40Bs).
• Maintain investment-grade balance sheet.
• Modest growing dividend ($0.60 annualized after the 30% March 2026 raise).
• Opportunistic buybacks ($650M in Q3 FY26).
• M&A only for adjacent capability.

Record net cash of +$24.4B (cash + investments $30.1B less total debt $5.72B, after a $4.4B debt paydown in Q3) plus ~$18B/quarter of FCF lets Micron self-fund the FY27 capex step-up to the mid-$40Bs without a raise, a balance-sheet position with no precedent in prior memory cycles, and the structural reason this cycle's drawdown risk is shallower than the historical -60% troughs.

ROIC / ROIIC by Cycle Phase

ROIIC (return on incremental invested capital) is the cleaner lens for HBM unit economics, incremental capex into HBM TSV packaging is returning well above 50% at current ASPs, lifting FY26E ROIC to record levels.

Working Capital & Cash Conversion

Inventory watch: Inventory of $8.6B is lean relative to the >$25B quarterly data-center run-rate, a sign of genuine shortage rather than channel build. The classic early-cycle warning is rising inventory days without revenue follow-through; currently the opposite, and ~$100B of SCA take-or-pay backlog floors demand visibility.

GAAP vs Non-GAAP Reconciliation (Q3 FY26)

The GAAP-to-non-GAAP gap is small, quality-of-earnings is excellent for a hardware company at this scale. Shares outstanding were 1,129M as of Jun 17, 2026 (the signed-off model uses ~1.1687B; the marketCap invariant holds at price × sharesOut).

3c · Historical Context: 10 Years Through 3 Cycles

10-Year Revenue & Gross Margin History

Cycle Comparison: Peak Magnitudes

The current cycle's peak revenue is ~3.5x the prior peak, with gross margin at all-time records (Q3 FY26 hit 84.9% vs prior ~62%). This is the structurally largest cycle in MU history, but every memory peak (1995/2000/2014/2018/2022) has been followed by a 55-90% drawdown. What's different this time: a record +$24.4B net-cash balance sheet and ~$100B of take-or-pay SCAs that floor roughly half of revenue.

Consensus EPS Revision Trajectory

Sell-Side Rating Changes (Last 90 Days)

Management Track Record

Sanjay Mehrotra (CEO since May 2017): co-founder and former CEO of SanDisk (sold to Western Digital for $19B in 2016, a textbook successful exit). Joined MU after 8 months "retirement". Reputation: disciplined operator, strong technologist (PhD electrical engineering), conservative on capex through troughs. Stock performance since his appointment: ~$28 → $1,132 (+3,940%). Survived the brutal 2018-19 and 2022-23 troughs with minimal balance-sheet damage. Notable strategic moves: pulled forward HBM3E investment in 2022, exited the 3D XPoint (Optane) joint venture cleanly, and secured $6.1B CHIPS Act funding for the Idaho/NY fabs. Buyback execution has been disciplined, bought aggressively at $50-60 in late 2023; MU has since crossed a $1T market cap (May 26, 2026) and a $1,255 52-week high.

Mark Murphy (CFO since 2018): ex-Qorvo CFO, conservative guidance posture (the $50B / $31 EPS Q4 guide is a notably large step-up for him), trusted by the buy-side. Manish Bhatia (EVP, Global Ops): runs the fabs; previously SanDisk. The team has been intact through this cycle, execution risk is low.

4 · Valuation

Current Multiples

5-Year Multiple History

Earnings Power

FY26 non-GAAP EPS is tracking ~$73 (Q1 $4.78 + Q2 $12.20 + Q3 $25.11 + Q4 guide $31), and FY27 consensus is ~$98.52 (+172% YoY). The stock trades at ~11.5x FY27E and ~25.5x TTM (on TTM EPS ~$44). Our $1,275 PT applies a 12.9x forward multiple, a modest re-rating off the ~11.5x current level, well below the AI cohort. The bear floor (~$620) prices a trough-normalized ~$62 EPS at ~10x; prior memory peaks unwound to low-single-digit-to-low-teens multiples on lower mid-cycle earnings.

Consensus

4b · Detailed Valuation: DCF + Comparable Companies

Segment-Level DCF Model

Below is a full segment-level DCF with 5-year explicit forecast and Gordon-growth terminal value. The four BUs project independently (revenue × EBITDA margin), aggregate to total, then we deduct CapEx (stepping to the mid-$40Bs in FY27), taxes (20% of EBITDA), and working-capital change (~2% of revenue change) to derive unlevered FCF. The DCF is a deliberately conservative cross-check: the $1,275 PT is multiples-anchored (12.9x FY27E EPS $98.52). Switch scenarios with the tabs, assumptions reload automatically. Sliders for WACC and terminal growth recompute the per-share NPV in real time.

Sensitivity Table: DCF NPV/Share ($) vs WACC × Terminal Growth

Reverse DCF: What Is the Market Pricing In?

The reverse-DCF tells us the market is pricing Micron less as a memory cyclical and more as an AI infrastructure franchise: implying that (a) the AI-capex cycle persists through 2030, (b) HBM becomes a multi-vendor share-stable oligopoly with Micron taking ~28%, and (c) terminal growth sits well above the semiconductor industry's GDP-plus norm. The ~$100B of take-or-pay SCAs make this thesis more defensible than in any prior cycle, but the terminal-state assumptions have never held through a memory down-cycle, which is exactly why we anchor the PT on multiples, not the perpetuity model.

Comparable Company Analysis

Full peer table, MU vs the two memory peers (SK Hynix, Samsung), the AI accelerator/infra cohort (NVDA, AVGO, AMD), the foundry leader (TSMC), the legacy IDM (INTC), and the SOXX sector ETF.

Mkt cap (signed-off MU shares 1.1687B) and growth as of June 2026. SK Hynix overtook Samsung's market cap on Jun 22, 2026; HBM share Q1 2026 SK Hynix 58% / Samsung 21% / Micron 21%. PEG = Fwd P/E ÷ Rev growth %. INTC shown for completeness, net loss makes P/E NM and PEG NM. Sources: StockAnalysis, GuruFocus, TrendForce/Counterpoint, MarketBeat.

Valuation Verdict

Our weighted view: We weight Bull 30% / Base 50% / Bear 20%, a blended fair value of ~$1,266, essentially at the $1,275 base PT. Net: Buy with discipline: own MU into the next print but begin trimming on the first sequential GM compression or a negative HBM4-share revision. Our $1,275 PT (12.9x FY27E EPS of $98.52) is in line with the ~$1,264 consensus mean and sits below the post-print cluster of $1,400-$2,000 targets, a modest, defensible re-rating rather than a stretch. At ~$1,132 spot, the stock offers ~+12.6% to target with a clearly contested Rubin HBM4 swing factor, which is why we maintain Buy rather than Strong Buy and counsel disciplined sizing.

5 · Upcoming Catalysts

Earnings Calendar

Conferences

• SEMICON West (Jul 2026), wafer/litho demand.
• Hot Chips (Aug 2026), Rubin/MI400 HBM4 detail; the cleanest read on flagship allocation.
• Citi Global TMT (Sep 2026).
• Ex-dividend (Jul 6, 2026), $0.15/qtr ($0.60 annualized).

Macro

• Fed: Cuts expected 2H 2026. Lower rates → more hyperscaler AI capex.
• China: No HBM restrictions beyond AI controls. CXMT trailing-edge DDR4/LPDDR4 could add 5-8% of global DRAM bit supply by FY28.
• Tariffs: Potential 100% tariff on imported memory, a windfall for MU as the only US producer.
• Sell-side PT migration: Post-print targets cluster $1,400-$2,000 (Barclays/DA Davidson/Susquehanna $2,000; Needham $1,650; TD Cowen/KeyCorp $1,600), continued upward revisions are a tailwind.

5b · Earnings Game Plan: Next Print: Q4 FY26 (Late Sept)

Q3 FY26: What Just Happened

Q3 FY26 (reported Jun 24, 2026) was a blowout: revenue $41.46B vs ~$35.8B consensus, non-GAAP EPS $25.11 vs a ~$20.5 whisper, a record 84.9% gross margin, and a Q4 guide to a record $50B / ~86% GM / $31 EPS. The stock popped ~13-15% on the print, set a $1,255 52-week high on Jun 25, then gave back -6.69% on Jun 26 to $1,132.33, a classic buy-the-rumor/sell-the-peak after a parabolic run. The playbook below now targets the Q4 FY26 print in late September.

Beat/Miss History: Last 12 Quarters

Q4 FY26 Setup

Implied Move & Volatility

Print-Day Playbook

Key Questions for the Call

• HBM4 Rubin allocation: "Do you hold flagship Vera Rubin HBM4 allocation, or is Micron's share concentrated in Rubin CPX (inference)? What's the multi-source split vs SK Hynix/Samsung?"
• HBM share: "With HBM4 in HVM at >$1B revenue, where do you see Micron's HBM share exiting CY26, holding ~21-23% or pushing toward 25%?"
• SCA coverage: "The 16 SCAs cover ~half of revenue, what's the path toward >50%, and how do the take-or-pay floors behave if DRAM ASPs reset?"
• FY27 capex: "The mid-$40Bs FY27 envelope is largely construction-related, how much is HBM4/HBM4E cleanroom vs greenfield, and what's the bit-supply implication?"
• DRAM ASP: "What's your view on conventional DRAM ASP trajectory through CY27 given HBM wafer cannibalization, and the risk of Samsung flooding once HBM4 qualifies?"
• Capital return: "At record +$24.4B net cash and ~$18B/qtr FCF, how should we think about buyback pace and the dividend trajectory?"

6 · Risk Factors

6b · Scenario Stress Tests: Quantified What-Ifs

The risks section frames the qualitative downside. Below we quantify each scenario, assuming it materializes in FY27, showing the EPS hit, revenue impact, and resulting price target. Each scenario carries an assigned probability that sums (with the base case) to 100%.

Probability-weighted EPS = $98.52 × 0.50 + $62 × 0.20 + $70 × 0.12 + $80 × 0.10 + $115 × 0.08 = $49.26 + $12.40 + $8.40 + $8.00 + $9.20 = ~$87.3 (vs base $98.52). The ~$100B SCA take-or-pay backlog provides a revenue floor absent in prior cycles, and +$24.4B net cash shallows the drawdown vs the historical -60% troughs.

Sensitivity to AI Capex (most consequential variable)

Cross-Scenario Conviction Check

• Rubin HBM4 to SK Hynix / MU to CPX (20% prob): The highest-conviction single-event downside. Watch TrendForce allocation updates, NVIDIA Vera Rubin BOM disclosures, and any flagship-vs-CPX confirmation on the Q4 call.
• Hyperscaler digestion (12% prob): Watch monthly Microsoft + Google + Amazon capex commentary, server motherboard order rates, and hyperscaler bond-issuance pace.
• Samsung DRAM flood (10% prob): Samsung began HBM4 shipments Feb 2026; once qualified, it has every incentive to deploy underutilized conventional-DRAM capacity. SCA price bands cushion but don't eliminate the reset.
• Flagship Rubin + tariff (8% prob): The tail-positive: a 100% memory tariff is a US-producer windfall, and flagship Rubin HBM4 allocation pushes FY27 EPS toward ~$115.

Hedging Implications

The probability-weighted EPS of ~$87.3 at a ~12-13x multiple implies ~$1,090-1,135, essentially the current ~$1,132 spot, with the published $1,275 PT carrying the modest re-rating credit. This is why we maintain a Buy but recommend disciplined sizing on strength: the asymmetry is positive (+12.6% to target, +48% bull) but bounded by the contested Rubin HBM4 question. A protective put spread at $1,000/$850 (cost ~2% of position) eliminates most of the downside in the bear scenario while preserving the upside to $1,680.

7 · Bull vs Bear Debate: Stress-Testing the View

Below is the strongest version of each side. We hold the Base case (the Pragmatist view) at 50%, the Bull's view at 30%, and the Bear's view at 20%. Reading both sides is essential before sizing a position.

Market-Implied Assumptions vs Our Three Scenarios

The reverse-DCF (Valuation section) already solved for what the market is pricing at $1,132. Here we lay those market-implied values against our base / bull / bear scenarios so the committee can see exactly where consensus sits:

Read: At ~$1,132 the market sits between the base and bull cases, pricing the base FY27 EPS at roughly an in-line multiple but not yet the full bull re-rating. Upside to $1,680 requires flagship Rubin HBM4 and ~$115 EPS; downside to $620 requires the SK-Hynix-led HBM4 lock-out plus a conventional-DRAM reset.

Counter-Argument to Each Bull Point

• "HBM is structural", but TSV equipment lead times are 18-24 months; constraint resolves by 2028 unless customers double-order again.
• "Hyperscaler capex secularly higher", true for 2026-27, but ROI scrutiny rises after the first AI-infra deployment cycle.
• "Industrial policy windfall", politically vulnerable to administration change; Korean lobbying is intense.
• "Multiple re-rating", has happened before (2017-18 cycle, 7x → 14x), then unwound to 5x within 24 months.
• "SCAs de-risk the cycle", take-or-pay price bands cushion volume but still float with ASPs; they soften, not eliminate, a reset.

8 · Ownership, Insiders & Flow

Short Interest & Float

Short interest is benign, below the 5% threshold that often signals contrarian setups; days-to-cover well under 1. MU crossed a $1T market cap for the first time on May 26, 2026, hit a $1,255 52-week high on Jun 25, then fell -6.69% on Jun 26 to $1,132.33. Short-interest and 13F specifics below are illustrative, verify live.

Options Snapshot

Implied move into the Q4 print (late Sept): ~±7% (~$80). Historical realized post-earnings: 9-11%. Options modestly underpricing, slight edge for a long straddle. Option strikes/IV here are illustrative, verify on a live chain.

Top Institutional Holders

Active managers (Capital, Fidelity, T. Rowe) are net buyers, a constructive signal. Passive ownership ~25% is normal for large-cap tech. Consensus rating is Strong Buy/Buy (38-43 analysts). 13F holder figures are illustrative, verify against the latest filings.

Insider Activity (Last 6 Months)

Read: Executive sales are all 10b5-1 pre-scheduled, typical hedging/diversification, not a signal. Director open-market purchases at $245-$280 (now $1,132, ~+304-362%) are the meaningful flag, directors don't often buy. Insider transactions here are illustrative, verify against SEC Form 4 filings.

9 · Technical Analysis

Analyst note: MU ran a parabolic year, from a $103 52-week low through month-end closes of $971 (May) to a $1,255 intraday 52-week high on Jun 25 (close $1,213.56) on the record Q3 print, then a -6.69% session on Jun 26 to $1,132.33. Primary uptrend intact (+259% YTD, +873% trailing year) but the structure is stretched and extended after the blowout, a classic sell-the-peak pullback. The next 5-10 sessions into the Q4 FY26 print (late Sept) resolve into continuation or distribution.

RSI/MACD/EMA levels are estimates, verify on your primary platform.

1 · Trend & EMAs

2 · Price Structure & Key Levels

3 · Chart Patterns

• Completed: Multi-leg breakout off the ~$103 52-week low, measured moves long since achieved.
• Active: Parabolic advance into the $1,255 print high; the -6.69% Jun-26 reversal is the first lower close.
• Forming: Potential blow-off-top digestion, a hold above $1,000 keeps the uptrend; a break below $971 opens the 50-EMA (~$760).
• Watch: Whether the Q3-print high ($1,255) caps a near-term range while earnings revisions catch up.

4 · Momentum

5 · Volume

• Up vs down volume: Up > down over the past 3 months, institutional accumulation footprint.
• Buying climax: The Q3-print pop to the $1,255 high came on heavy volume, exhaustion behavior.
• Reversal volume: The -6.69% Jun-26 session was a distribution day, watch follow-through.
• OBV: New highs into the print; watch for a failure on the next rally as a distribution tell.

6 · Relative Strength · Multi-Timeframe Peer Performance

7 · Multi-Timeframe Summary

8 · Trade Setups

Technical bottom line: Fundamentals are record-strong but the chart is parabolic and just reversed -6.69% off the print high. Wait for a $1,000-1,010 retest (preferred) or a high-volume reclaim of $1,255. Long-term uptrend intact above the $971 May close.

9b · Trading Toolkit: Options, Sizing, Credit

Options Strategy Ideas (IV elevated post-print)

Strikes/premia below are illustrative against a ~$1,132 spot, verify on a live chain.

Position Sizing Framework

Credit / Bond View

With a record +$24.4B net-cash position (total debt cut to $5.72B after a $4.4B Q3 paydown) and ~$18B/qtr of FCF, MU's credit is the strongest in its history, CDS spreads signal no distress, and an upgrade is plausible if management holds this capital structure through the cycle. Specific agency ratings/CDS levels are directionally current but were not re-pulled this pass, verify live.

★ Interactive Models & Calculators

Ask the Thesis AI-assisted checking…

Describe a scenario in natural language; the assistant returns a structured impact analysis against this dashboard's DCF, HBM share assumptions, and scenario math. Powered by Claude via a Cloudflare Worker proxy (Anthropic key held server-side; same pattern as the live-quote feed).

Note: The assistant reasons from the dashboard's data snapshot and thesis sections, it does not browse the web, hit Bloomberg, or access real-time fundamentals beyond what's in data.js. Treat its responses as scenario-modeling support, not as primary research. Author judgments on rating, PT, and probabilities remain with the analyst.

Glossary & Methodology Notes

Memory Technology Glossary

Financial / Valuation Terms

Methodology Notes

• DCF model: Explicit forecast FY26-FY30; Gordon-growth terminal. Segment-level revenue × EBITDA margin → unlevered FCF after capex, taxes (20%), and working capital. Discount at WACC. The DCF is a conservative cross-check; the $1,275 PT is multiples-anchored (12.9x FY27E EPS $98.52), cross-checked with an EV/EBITDA bridge.
• Comp set: Memory peers (SK Hynix, Samsung) for direct comparison; AI infra (NVDA, AVGO, AMD) for sentiment context; TSMC for foundry contrast; INTC for legacy IDM contrast; SOXX as sector benchmark.
• Probability weighting: Bull 30% / Base 50% / Bear 20% → blended ~$1,266, essentially at the $1,275 base PT. The bear weight is lifted from a prior 15% because the contested NVIDIA Vera Rubin HBM4 flagship-vs-CPX allocation question (SK-Hynix-skewed) is a genuine downside swing.
• Technical levels: EMA/RSI/MACD values are estimates against the monthly close series (52-week range $103.38-$1,255.00); actual values may differ. Verify on a live platform.
• Data freshness: Operational data (revenue, EPS, guidance) is current as of the Q3 FY26 release (quarter ended May 28, 2026; reported Jun 24, 2026). Pricing data as of the June 26, 2026 close ($1,132.33).
• Share count: The model uses ~1.1687B shares (signed-off) to preserve marketCap = price × sharesOut; the Q3 FY26 10-Q shows 1,129M shares outstanding (Jun 17, 2026) and 1,145M diluted weighted-average. Live market cap on the current count is ~$1.28T.

Recent Corrections (Verification Audit)

• Full rebuild for Q3 FY26 (Jun 26, 2026): The dashboard was rebuilt from a stale snapshot. Price $1,020 → $1,132.33 (Jun 26 close, -6.69% on day); YTD +223.5% → +259.2%; 52-week range $90.93-$1,110.40 → $103.38-$1,255.00; market cap $1,192B → $1,323B (= $1,132.33 × 1.1687B); fwd P/E 11.4x → 11.5x on FY27E EPS $89.58 → $98.52.
• Rating & PT: Buy maintained; 12-month PT $750 → $1,275 (12.9x FY27E EPS $98.52, ~$1,264 consensus mean); bull $1,000 → $1,680; bear $360 → $620; probability weights 35/50/15 → 30/50/20.
• Q3 FY26 results: Replaced the old Q2/Q3-guide figures with the reported record quarter, revenue $41.46B (+346% YoY), non-GAAP GM 84.9%, non-GAAP EPS $25.11, and the Q4 guide ($50.0B / ~86% GM / $31.00 EPS). Business units, product mix (DRAM $31.3B / NAND $9.9B), and all charts rebased.
• Net cash: Updated from +$4.5B to a record +$24.4B (cash + investments $30.1B less total debt $5.72B after a $4.4B Q3 paydown); the DCF net-cash constant was changed accordingly.
• HBM: Share refreshed to Q1 2026 (SK Hynix 58% / Micron 21% / Samsung 21%). HBM4 is now in high-volume shipments (>$1B revenue, 12-high ramp ~2x HBM3E). The central new risk is the NVIDIA Vera Rubin HBM4 allocation skewing to SK Hynix, with Micron potentially confined to inference-tier Rubin CPX, the reason the bear weight rose to 20%.
• SCAs: Added the 16 Strategic Customer Agreements ≈ $100B take-or-pay backlog covering ~half of revenue, the structural cycle de-risk that distinguishes this cycle from prior boom-busts.
• Capex & dividend: FY26 capex ~$27B; FY27 stepping to the mid-$40Bs. Dividend raised 30% in Mar 2026 to $0.60 annualized; $650M buyback in Q3.

Author's positions & verification note