MU — Earnings Review Q1 FY26 (Atlas)

Date: 2026-02-23 Quarter: Q1 FY26 (September–November 2025, reported December 17, 2025) Market cap: ~$108B | P/S TTM: ~2.6x | Revenue growth: +56.7% YoY | EPS growth: +167% YoY

Verdict

Q1 FY26 was Micron's best quarter in its 47-year history on every metric simultaneously: revenue, gross margin, operating margin, FCF, EPS — all all-time highs. The $13.6B print beat guidance midpoint by $1.1B (+9.1%). Q2 FY26 guidance of $18.7B implies another 37% sequential surge, with 68% gross margin guiding 11pp above Q1's record. HBM is sold out through calendar 2026 under locked LTAs. The valuation — 2.6x trailing P/S, ~5.2x trailing P/E, ~14% FCF yield — is the market demanding proof this cycle doesn't reverse. Conviction 4/5: extraordinary execution and historic cheapness argue for 5, but secular-versus-cyclical resolution is Q3/Q4 FY26's job.

Qualification Gate

Criterion Threshold Actual Pass?
Revenue YoY growth >30% (>40% preferred) +56.7% PASS
Gross margin >60% preferred 56.0% GAAP / 56.8% Non-GAAP (68% guided Q2) BORDERLINE — improving
Revenue per quarter >$50M $13.6B PASS
Data availability 4+ quarters 37 quarters in DB PASS
Share dilution <10% annual ~1% annual PASS
GAAP profitability trajectory Improving $0 → $5.2B net income, 0% → 38.4% NM in 6 quarters PASS

Gross margin is the only soft point. GAAP GM of 56% misses the 60% threshold on a trailing basis. Q2 FY26 guide implies 68% — well above threshold. Not a disqualifying failure given cyclical context and clear trajectory.

Six-Factor Score

Factor Rating Detail
Growth Strong +56.7% YoY Q1 FY26; DRAM +69% YoY. Three consecutive acceleration quarters.
Trajectory Accelerating +36.6% Q3 FY25 → +46.0% Q4 FY25 → +56.7% Q1 FY26 → Q2 guide implies ~79% YoY
Margins Mid / Rapidly improving GAAP GM 56% (ATH); Non-GAAP GM 56.8% (ATH); 68% guided Q2. OM 45% GAAP / 47% Non-GAAP (ATH).
Dominance Strong #2 DRAM globally (25.7% share). #2 HBM globally (21% share, overtaking Samsung 17%). HBM TAM now concentrated: SK Hynix 62%, Micron 21%, Samsung 17%.
Valuation Cheap 2.6x P/S TTM, ~5.2x trailing P/E, ~14% FCF yield. At Q2 guide annualized: ~1.5x P/S, ~2.8x P/E. Extraordinary cheapness if margins hold even partially.
Special Present HBM TAM pulled forward 2 years: $35B (2025) → $100B (2028). Supply sold out CY26 under LTAs. 50-67% of hyperscaler demand currently unmet. Technology lead: HBM3E 30% lower power than SK Hynix; HBM4 at 11+ Gbps (industry-leading).

The Numbers (12 Quarters)

Q2 FY23 Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Q4 FY24 Q1 FY25 Q2 FY25 Q3 FY25 Q4 FY25 Q1 FY26
Cal period Mar-2023 Jun-2023 Sep-2023 Dec-2023 Mar-2024 Jun-2024 Sep-2024 Dec-2024 Mar-2025 Jun-2025 Sep-2025 Dec-2025
Revenue ($B) 3.75 3.76 4.01 5.82 6.81 7.75 7.75 8.71 8.05 9.30 11.30 13.60
YoY% -53% -57% -40% +18% +82% +82% +93% +84% +18% +20% +46% +57%
QoQ% -10% 0% +7% +45% +17% +14% 0% +12% -8% +16% +22% +21%
GAAP GM% -33% -15% +12% +22% +28% +33% +35% +39% +39% +39% +47% +56%
GAAP OM% -62% -44% -18% -6% +4% +12% +12% +25% +25% +26% +34% +45%
GAAP NM% -63% -45% -18% -7% +5% +13% +13% +25% +25% +26% +35% +38%
GAAP EPS ($) -1.73 -1.43 -0.41 0.45 0.62 1.18 1.18 1.79 1.56 1.91 3.05 4.48
Non-GAAP EPS ($) 4.78
FCF ($B) 1.3 0.9 1.9 3.0 3.9

Note: Historical quarters Q2 FY23–Q4 FY24 use EDGAR-sourced data from DB. Non-GAAP EPS prior quarters not reconstructed in DB. FCF Q1-Q3 FY25 estimated from operating CF minus CapEx.

Q2 FY26 Guide: Revenue 18.7B±400M (+37% QoQ, ~79% YoY) | GAAP GM ~67-68% | Non-GAAP EPS 8.42±0.20

Prior Beliefs → Updated Beliefs

Entering Q1 FY26 from my Q3 FY25 analysis, expecting continuation of HBM ramp with guided beat.

Metric Expected (Prior) Actual Verdict
Revenue $12.5–13.0B (guided $12.5B ± $300M, expected modest beat) 13.6B(+1.1B above guide midpoint) Beat — guided beat magnitude again exceeded
GAAP Gross Margin ~50–52% (extrapolating from $47% Q4 FY25) 56.0% GAAP / 56.8% Non-GAAP Significant upside surprise — pricing power beyond model
GAAP EPS ~$3.50–4.00 $4.48 GAAP / $4.78 Non-GAAP Beat — margin flow-through underestimated
FCF ~$2.5–3.0B $3.9B (quarterly record, +386% QoQ) Beat — cash conversion exceptional
HBM status Sold out CY25, partial CY26 sold 100% CY26 sold out under LTA; "50–67% of demand" unmet Stronger than expected — LTA coverage confirmed
Forward guidance $15–16B Q2 $18.7B — blew past expectation Massive upside; implies HBM volume ramp steeper than modeled
Gross margin trajectory ~58–60% Q2 68% Non-GAAP guided Extraordinary — 12pp sequential GM expansion guided

Delta: Every single line beat. The guidance beat surprise is the most important number — 18.7BQ2guidewas 3B above the sell-side consensus. The GM guide of 68% implies DRAM/HBM ASP expansion continues even as volumes surge, which is the unusual dynamic of supply-constrained AI memory. This is not a normal semiconductor cycle.

Leading Indicators

Bullish divergence — all indicators accelerating together:

Indicator Q3 FY25 Q4 FY25 Q1 FY26 Trend
HBM capacity sold CY25 sold out CY26 partial CY26 100% locked Accelerating — supply constraint extends timeline
DRAM ASP +sequential +mid-teens +20% Acceleration
NAND ASP +sequential n/a +mid-teens Stable positive
Data center revenue % ~40% ~50% ~54% ($7.7B) Secular shift, mix positive
Cloud Memory BU n/a n/a $5.3B (39% of rev) New ATH
Operating cash flow $8.4B Highest ever
FCF $1.9B $3.0B $3.9B +105% in two quarters
Inventory days ~140+ ~130 126 days Declining (normalization)
CapEx guidance ~$14B ~$18B $20B FY26 Raised — conviction in demand durability

No divergence visible. Revenue, pricing, FCF, and forward guidance are all pointing in the same direction. The only watch item: inventory at 126 days is still elevated versus the ~90-day normalized range, signaling some caution on NAND/commodity DRAM even as HBM demand far outstrips supply.

Q2 FY26 guide math: $18.7B / $13.6B = +37% QoQ. At 68% GM, gross profit would be $12.7B vs Q1's $7.6B — a $5.1B QoQ gross profit expansion in a single quarter. If accurate, FCF could approach $6–7B for Q2. That would put annualized FCF at $25–28B against a $108B market cap, yielding 23–26% FCF yield.

Scuttlebutt Findings

From stages/scuttlebutt/MU/2026-02-23.md:

Valuation Context

Metric Current 1Y Ago Assessment
Market cap ~$108B ~$100B Roughly flat despite earnings tripling
P/S TTM ~2.6x ~3.5x Compression — market pricing in cycle peak
P/S Fwd (Q2 ann.) ~1.5x Extraordinarily cheap if $18.7B guide delivers
P/E trailing ~5.2x (on Q1 record) n/a (loss) Historical low; price has barely moved vs earnings
P/E fwd (Q2 ann.) ~2.8x The market is pricing in a 70–80% earnings reversal
FCF yield TTM ~14% High single-digit to mid-teens; rising fast
FCF yield fwd (Q2 ann.) ~23–26% If Q2 sustains; implies market prices in sharp reversal H2 FY26

Peer comparison: SK Hynix trades at ~6–8x forward P/E. Samsung at ~15x (blended). NVDA at 35–40x forward P/E (the demand driver). Micron trades at a deep discount to both its direct peers and the AI stack it enables. The discount reflects its commodity-cyclical history. The question is whether HBM changes the underlying characterization.

What the market is saying: At $108B market cap and a $33.68 annualized EPS run rate (at Q2 guide), the market is implying FY27/FY28 earnings will revert significantly toward the $3–5B range (implying a 10–15x forward P/E at trough). This is the embedded bear case. If HBM LTAs hold pricing through FY27, the stock is dramatically mispriced.

Platform & Secular Position

Micron is not a platform company. It is a commodity-to-specialized memory manufacturer transitioning its mix toward specialized, high-margin memory (HBM) for AI infrastructure.

Secular tailwind: AI compute = GPU memory bandwidth bottleneck. HBM is the only solution at scale. TAM: $35B (CY25) → $100B (CY28) — 40% CAGR, pulled forward 2 years vs. prior estimates. Micron's 21% share of a $100B market = 21BHBMrevenuebyCY28vs6B estimated FY26. The flow-through at 65%+ GM would be transformative.

The structural argument: Prior memory cycles were characterized by interchangeable supply (any DRAM is a DRAM). HBM breaks this. It requires dedicated technology (TSV stacking, thermal management, power efficiency), 2-3 year manufacturing lead times, and hyperscaler qualification. Supply cannot respond quickly to price signals. LTAs further lock pricing/volume multi-quarter out. This is why Mehrotra says "50–67% of demand currently unmet" while simultaneously calling CY26 sold out — there is structural shortage, not cyclical shortage.

Counterargument: Commodity DRAM and NAND remain 40–50% of revenue. These do cycle. A consumer/enterprise PC refresh that disappoints (possible given macro) pressures non-HBM pricing. Samsung remains a latent supply threat — they are incentivized to regain HBM share and have enormous capacity. If Samsung's HBM3E qualifies at NVIDIA in CY26, HBM pricing would come under pressure. Mehrotra's "50–67% demand unmet" claim has not been independently verified.

Key Risks

  1. Cyclical mean-reversion — Memory has reverted every prior cycle. If hyperscaler GPU buildout plateaus in CY26-27 (due to ROI concerns on AI capex), HBM demand softens and LTA prices come under renegotiation pressure. Commodity DRAM/NAND could correct simultaneously.

  2. Samsung HBM qualification — Samsung is actively working to qualify HBM3E at NVIDIA and other hyperscalers. If successful, HBM supply increases materially without corresponding demand growth, compressing pricing. Timeline uncertain — could be H2 CY26 or CY27.

  3. Customer concentration — Four hyperscalers drive the majority of HBM revenue. If one reduces AI capex plans (any of Microsoft, Google, Amazon, Meta), Micron feels it disproportionately. No disclosed customer represents >10% of revenue (public disclosure), but hyperscaler concentration in HBM is real.

  4. CapEx execution and stranded assets — FY26 CapEx raised to $20B. $200B committed to US fabs over decade. If demand deteriorates before Idaho Fab 1 (mid-CY2027) comes online, CapEx outspends FCF and balance sheet deteriorates. Prior cycle: Micron's debt peaked at $13.7B (now $11.8B declining).

  5. Tariff / geopolitical risk — China trade tensions (300+ job cuts in mobile NAND already). Any escalation impacting China sales or supply chain disrupts ~10% of revenue. CHIPS Act subsidies are a partial offset but politically contingent.

Key Catalysts

  1. Q2 FY26 earnings (March 2026) — The $18.7B / 68% GM test. Delivery validates HBM LTA thesis and forces market re-rating. A beat is the single highest-probability near-term catalyst. Guidance beat pattern: +7% Q3 FY25, +27% Q4 FY25, +9% Q1 FY26 — three consecutive beats.

  2. LTA disclosures — Mehrotra called CY26 HBM agreements "much stronger" in structure vs prior but deferred specifics. Formal LTA announcements with volume/price commitments from hyperscalers would provide multi-year revenue visibility and force SOTP revaluation.

  3. HBM4 volume ramp (H2 CY26) — HBM4 at 11+ Gbps, higher ASPs than HBM3E. Qualification at major hyperscalers for next-generation AI accelerators. Successful ramp extends the technology lead vs Samsung and sustains GM expansion.

  4. Samsung disqualification confirmation — If Samsung fails to qualify HBM3E at NVIDIA into CY26, HBM supply remains a two-player (SK Hynix + Micron) market through CY27. Ongoing Samsung struggles at qualification = Micron pricing power sustained.

  5. Idaho Fab 1 pull-forward — First wafers now expected mid-CY2027 (pulled in from prior estimates). US domestic production qualifies for CHIPS Act subsidies and addresses hyperscaler supply-chain resilience requirements. Formal subsidy award announcement would be a catalyst.


Conviction: 4/5. Position: Not disclosed.