MU · Q4 2025 Earnings

Bullish

Micron

Reported September 23, 2025

30-second summary

Micron closed FY25 with $11.32B in Q4 revenue (+46% YoY, +22% QoQ), GAAP gross margin of 44.7% — +270bps above the GAAP guide high end — and Cloud Memory revenue more than tripling YoY to $4.54B. The Q1 FY26 guide of $12.5B at the midpoint implies another ~$1.2B sequential step, with GAAP gross margin guided to 50.5% (±100bps) — the first time Micron has guided north of 50% in this cycle. CapEx steps from $13.8B (net, FY25) to ~$18B (net, FY26), almost entirely DRAM, signalling management is now spending against a structural HBM/leading-edge DRAM demand thesis rather than a cyclical recovery.

Headline numbers

EPS

Q4 FY2025

$3.03

Revenue

Q4 FY2025

$11.32B

+46.1% YoY

Gross margin

Q4 FY2025

44.7%

Operating margin

Q4 FY2025

32.3%

Key financials

Q4 FY2025

Metric	Q4 FY2025	YoY	Q3 FY2025	QoQ
Revenue	$11.32B	+46.1%	$9.30B	+21.7%
EPS	$3.03	—	$1.91	+58.6%
Gross margin	44.7%	—	39.0%	+570bps
Operating margin	32.3%	—	26.8%	+550bps

Guidance

Guidance is issued one quarter forward. The Prior-guide column references the guide issued last quarter for the period just reported; the New-guide column is for next quarter.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q4 FY2025	$10.7 billion ± $300 million	$11.32 billion	+$0.62 billion above guide high	Beat
EPS (GAAP)	Q4 FY2025	$2.29 ± $0.15	$2.83	+$0.54 above guide high	Beat
EPS (Non-GAAP)	Q4 FY2025	$2.50 ± $0.15	$3.03	+$0.53 above guide high	Beat
Gross Margin (GAAP)	Q4 FY2025	41.0% ± 1.0%	44.7%	+3.7 percentage points above guide high	Beat
Gross Margin (Non-GAAP)	Q4 FY2025	not guided	45.7%	—	Beat
Operating Expenses (GAAP)	Q4 FY2025	$1.35 billion ± $20 million	not separately disclosed	—	Beat

New guidance

Metric	Period	Guide	YoY
Revenue	Q1 FY2026	$12.5 billion ± $300 million	—
EPS (GAAP)	Q1 FY2026	$3.56 ± $0.15	—
EPS (Non-GAAP)	Q1 FY2026	$3.75 ± $0.15	—
Gross Margin (GAAP)	Q1 FY2026	50.5% ± 1.0%	—
Gross Margin (Non-GAAP)	Q1 FY2026	51.5% ± 1.0%	—
Operating Expenses (GAAP)	Q1 FY2026	$1.49 billion ± $20 million	—

Segment performance

Q4 FY2025

Segment	Q4 FY2025	YoY
Cloud Memory Business Unit	$4.54B	+213.3%
Mobile and Client Business Unit	$3.76B	+24.5%
Automotive and Embedded Business Unit	$1.43B	+16.7%
Core Data Center Business Unit	$1.58B	-23.0%

Profitability

Q4 FY2025

Segment	Q4 FY2025
Cloud Memory Business Unit Gross Margin	59%
Cloud Memory Business Unit Operating Margin	48%
Mobile and Client Business Unit Gross Margin	36%
Mobile and Client Business Unit Operating Margin	29%
Core Data Center Business Unit Gross Margin	41%
Automotive and Embedded Business Unit Gross Margin	31%
Operating Cash Flow	$5.73B
Non-GAAP Gross Margin	45.7%

Management tone

Narrative arc: Q3 HBM scaling faster than planned → Q4 Cloud memory at scale, building for the next leg.

Last quarter management was still arguing that HBM share parity could be pulled forward to H2 CY25; this quarter the conversation has moved past parity to capital allocation for what comes after. The ~30% CapEx step from $13.8B to ~$18B (net, FY26), almost entirely directed at DRAM construction and equipment, is the most concrete signal of management's confidence in multi-year demand. Sanjay's framing in Q&A — that the new Idaho fab won't produce meaningfully until H2 2027 — implicitly underwrites a tight supply environment through at least CY26.

On HBM specifically, the posture continued to harden. In Q3 management characterized CY26 HBM bit demand as "significantly exceeding" overall DRAM industry growth while stopping short of declaring CY26 booked; in Q4 Q&A, management told Morgan Stanley that HBM share in CY26 will exceed CY25 levels and explicitly framed HBM as a "separate business model with long lead times, locked-in volumes and pricing." That language — locked-in — is meaningfully more committal than the Q3 process-oriented framing on CY26 customer conversations.

The margin story shifted from cyclical recovery to structural sustainability. To Citi, management enumerated seven distinct margin drivers: continued price increases, tight DRAM, improved NAND, higher-value SSDs, sub-target inventory, extended DDR4/LP4 life constraining tech transitions, and HBM silicon intensity. The phrase that operating margin is now the highest since November 2018 anchors how far the cycle has travelled — but the supply-side levers (fab ramp timing, DDR4 extension, HBM silicon intensity) are being framed as enduring, not transient.

Notable absence: the "memory market equilibrium is being gradually restored" register from Q3 has dropped out. Q4 prepared remarks and Q&A read as a company that has stopped hedging against a normalization scenario.

Q&A highlights

Tom O'Malley · Barclays

Analyst seeks clarity on NAND market conditions, hyperscaler demand for storage, capacity additions, and CapEx guidance breakdown between DRAM and NAND spending.

Management characterized Q4 NAND bit declines as mix-driven noise. Expects NAND industry improvement in calendar 2026 driven by hyperscaler AI server deployments requiring more storage due to HDD shortages. Confirmed CapEx increasing from $13.8B (net, FY25) to ~$18B (net, FY26), with vast majority directed to DRAM construction and equipment.

FY25 net CapEx: $13.8B; FY26 net CapEx: ~$18BVast majority of incremental CapEx for DRAM construction and equipmentStructurally brought down NAND wafer outs and slowing node transitionsData center SSD market share hitting record levels

Joe Moore · Morgan Stanley

Analyst asks about HBM share trajectory in 2026, sustainability of share gains in HBM4, and whether HBM gross margins will remain higher than non-HBM DRAM given pricing lock-in versus spot DDR5 market.

Management expects HBM share to exceed FY25 levels in FY26. Characterized HBM as separate business model with long lead times, locked-in volumes and pricing, and higher through-cycle ROI. Doesn't view lock-in negatively; notes potential for non-HBM DRAM to match or exceed HBM profitability at times when industry is healthy.

Expected to reach DRAM supply share vicinity by calendar Q3 2025HBM treated as separate business with different ROI profileLong lead times and locked-in pricing provide stable returnsManagement comfortable with HBM share shifts between competitors if overall share maintained

Chris Dainley · Citi

Analyst seeks DRAM mix breakdown (DDR4/LP4 versus DDR5 versus HBM) and relative margin contributions; asks what changed to enable gross margins above 50% (previously cited as difficult) and identifies limiting factors.

DDR4 (excluding LP4) is low single-digit percentage of revenue; extended EOL timelines due to shortages have improved margins. Gross margin improvement driven by continued price increases, tight DRAM market, improved NAND market, higher-value SSD products, low inventory levels, extended DDR4/LP4 life constraining tech transitions, HBM silicon intensity growth, and limited incremental capacity. Second quarter gross margin expected to improve sequentially.

DDR4 now low single-digit % of business (previously cited as high single-digit)Margin above mid-FY22 levels; operating margin highest since November 2018DDR4/LP4 extended life constraining technology transitionsLow inventory below targets supporting pricing power

Vijay Rakesh · Mizuho

Analyst asks about 1-gamma node ramp trajectory versus 1-beta, expected mix progression through FY26, and HBM4 margin expectations versus HBM3E.

1-gamma achieved mature yields and first hyperscaler revenue shipments in Q4. Expected to be primary growth driver in FY26 across multiple product areas. 1-gamma plus 1-beta already represent significant majority of bit output, with mix continuing to shift toward 1-gamma through year. HBM4 costs higher than HBM3E; prices will be meaningfully higher. Management confident in HBM4 ROI but doesn't disclose specific margin comparisons by product line.

1-gamma achieved mature yield and first revenue shipments in Q4 FY251-gamma + 1-beta = significant majority of bit output currently1-gamma to be primary growth driver in FY26HBM4 pricing meaningfully higher than HBM3E pricing

Kevin Cassidy · Rosenblatt Securities

Analyst asks about long-term supply agreements with customers and rationale for exiting managed NAND smartphone market.

Management noted customer interest in long-term agreements but emphasized need for thoughtfulness given U.S. manufacturing ramps, potential tariff changes (Section 232), and strategic mix shift toward data center. Exited managed NAND because mobile market ROI insufficient; preferred to concentrate resources on higher-ROI data center SSD business. Customers transitioned but DRAM relationship remains critical for them.

Exiting managed NAND due to low ROI in mobile segmentMobile NAND pricing expectations and competition incompatible with robust ROISecond year of positive free cash flow in NAND reportedDRAM business with smartphone makers remains critical despite NAND exit

Answers to last quarter's watch list

Q4 non-GAAP GM at or above 42% guide; commitment to expansion into Q1 FY26 — Non-GAAP gross margin came in at 45.7%, 370bps above the guide midpoint, and Q1 FY26 non-GAAP GM is guided to 51.5% — a ~580bps sequential expansion.

Resolved positively

Explicit CY26 HBM booking visibility or pricing terms — Management told Morgan Stanley HBM is a "separate business model with long lead times, locked-in volumes and pricing" and confirmed CY26 HBM share will exceed CY25 — meaningfully more committal than Q3's process-oriented language. Specific CY26 booking percentage was not disclosed.

Resolved positively

HBM share trajectory vs H2 CY25 parity target — Management confirmed expectation to reach DRAM supply share vicinity by calendar Q3 2025 in Q&A, with CY26 HBM share to exceed CY25 levels. Specific Q4 HBM revenue dollar figure not broken out (Cloud Memory BU at $4.54B is the closest disclosure). Status: Resolved positively, though disclosure granularity reduced.

DRAM pricing direction for Q1 FY26 — Management cited "continued price increases" as a Q1 margin driver and pointed to a tight DRAM market with low inventory below targets. Implied pricing direction is up.

Resolved positively

NAND underutilization resolution — Management confirmed NAND wafer outs have been "structurally" brought down and node transitions slowed — i.e., the underutilized capacity is being permanently removed, not reactivated. Calendar 2026 NAND demand expected to improve via hyperscaler SSD pull.

Resolved positively

Singapore/Idaho capacity expansion timing — Idaho fab will not produce meaningful output until H2 2027, per management's Q&A response to Citi. No pull-forward. FY26 CapEx steps to ~$18B (net) from $13.8B, with vast majority to DRAM — but most of that spend is for capacity arriving in CY27+.

Continue monitoring

What to watch into next quarter

Whether Q1 FY26 GAAP gross margin lands at or above the 50.5% guide midpoint — and whether management guides Q2 FY26 above 52%. The current guide already prices in significant structural margin; a Q2 deceleration in guided GM would be the first crack in the structural thesis.

Disclosure on Core Data Center BU -23% YoY — confirm whether this is reporting reclassification (Cloud SSDs/HBM moving to Cloud Memory) or genuine end-market weakness. The opacity here is the one place the segment story doesn't add up cleanly.

HBM4 ramp cadence and any explicit margin commentary by product — management declined to give product-line margins this quarter. A change in disclosure posture would signal either confidence (likely positive) or pressure to reset expectations.

Quantified CY26 HBM supply commitments — "locked-in volumes and pricing" was the Q4 verbal upgrade; watch for a percentage-booked figure or a dollar commitment figure to match the Q3 "CY25 sold out" precedent.

FY26 CapEx allocation detail — $18B net is the headline; watch whether subsequent quarters reveal the DRAM/HBM/Idaho construction split, and whether any of the spend is pulled forward into FY26 production (vs the stated H2 2027 Idaho start).

DDR4/LP4 extension dynamics — now low single-digit % of revenue. If customers begin transitioning faster than expected, the constraint on technology node transitions (cited as a margin support) weakens.

Sources

Micron Technology, Inc. Fiscal Q4 2025 Press Release (Form 8-K Exhibit 99.1), filed September 23, 2025. https://www.sec.gov/Archives/edgar/data/723125/000072312525000024/a2025q4ex991-pressrelease.htm
Micron Technology Fiscal Q4 2025 earnings call Q&A (analyst exchanges as transcribed).

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.