Research — Oct. 27, 2026
Behind the Earnings Call: Micron Technology Inc. Q4 2026
By Janya Goel
Overview
Micron closed fiscal 2026 on a high. Record revenue, gross margin and EPS in Q4 all came in above the top of its guidance. CEO Sanjay Mehrotra called the year "outstanding," with revenue 3.5x last year's record, data center revenue up fourfold, and DRAM revenue passing $100 billion.
Financial performance and consensus
Powered by Visible Alpha Estimates
- Q4 revenue was a record $54.2B (sixth straight record, 6.4% above Visible Alpha consensus; FY2026 $133.2B), with 87% gross margin, non-GAAP EPS of $33.42 (4.9% above expectations), DRAM and NAND ASP growth of 19.2% and 28.7% sequentially (vs. consensus 17.4% and 19.5%), $33.2B free cash flow and two credit rating upgrades.
- Consensus reaction: Visible Alpha raised Q1 FY2027 revenue expectations by 7.7% and EPS by 7%, on stronger DRAM and NAND pricing.
- FY2027 consensus and costs: Revenue is expected to more than double to $277.4B, and net income to rise 131% to $196B. OpEx should rise by about $2.5B over the year, driven by R&D and incentive compensation, and the tax rate should be about 15.5%.
- CapEx and capital return: CapEx steps up to about $11.5B in Q1 and about $25B in 1H, with 2H higher, and more is going to new clean rooms than to equipment. Micron plans to return more cash starting December 9, mainly through buybacks, and expects to seek a larger authorization than the current $2.2B.
Micron: Visible Alpha consensus expectations, Q4 2026 earnings surprise, and forward revisions
Key measures (Operating) |
Surprise — Q4 2026 |
Revisions — Q1 2027e |
2026 |
2027e |
2028e |
||
PreQ estimates |
Actuals |
PreQ estimates |
Actuals |
||||
Revenue ($M) |
50,952 |
54,229 ↑ |
57,568 |
61,979 ↑ |
133,188 |
277,375 |
326,897 |
Gross margin (%) |
86.5 |
87.0 ↑ |
87.0 |
86.5 ↓ |
81.1 |
87.2 |
86.0 |
Net income ($M) |
36,115 |
37,701 ↑ |
41,018 |
43,676 ↑ |
84,969 |
196,010 |
228,575 |
Diluted EPS to common stock ($) |
31.30 |
32.87 ↑ |
35.54 |
38.04 ↑ |
74.34 |
171.87 |
203.93 |
DRAM segment breakdown |
|||||||
DRAM revenue ($M) |
38,454 |
39,771 ↑ |
43,056 |
45,106 ↑ |
100,679 |
202,669 |
246,536 |
DRAM ASP - sequential growth (%) |
17.4 |
19.2 ↑ |
7.4 |
8.5 ↑ |
178.2 |
69.8 |
(1.3) |
DRAM gross profit ($M) |
34,210 |
35,364 ↑ |
38,468 |
40,326 ↑ |
84,212 |
184,350 |
218,939 |
Gross margin - DRAM (%) |
88.2 |
88.8 ↑ |
88.8 |
88.4 ↓ |
83.6 |
88.9 |
88.1 |
NAND segment breakdown |
|||||||
NAND revenue ($M) |
12,282 |
14,102 ↑ |
13,822 |
15,942 ↑ |
31,785 |
69,132 |
79,124 |
NAND ASP - sequential growth (%) |
19.5 |
28.7 ↑ |
7.9 |
8.7 ↑ |
203.5 |
78.5 |
(7.4) |
NAND gross profit ($M) |
10,282 |
11,772 ↑ |
11,712 |
13,411 ↑ |
23,688 |
59,228 |
64,586 |
Gross margin - NAND (%) |
82.7 |
83.6 ↑ |
83.8 |
83.1 ↓ |
74.6 |
83.7 |
82.8 |
Operating cash flow ($M) |
32,720 |
43,973 ↑ |
39,970 |
39,940 ↓ |
89,675 |
195,166 |
243,762 |
Total operating expenses - operating ($M) |
1,739 |
2,568 ↑ |
1,808 |
2,110 ↑ |
6,841 |
9,584 |
11,261 |
Capital expenditure ($M) |
10,623 |
11,110 ↑ |
12,195 |
11,929 ↓ |
30,712 |
56,194 |
67,540 |
Data compiled Oct. 8, 2026. e = estimate; PreQ = prequarter; EPS = earnings per share. Arrow reflects the change in current vs. PreQ expectations. Tan (↑): current estimates higher than PreQ; Gray (↓): current estimates lower than PreQ. Source: Visible Alpha, a part of S&P Global Market Intelligence. © 2026 S&P Global.
What the earnings call revealed
Powered by ProntoNLP Transcript Analytics
ProntoNLP surfaces business event scores, sentiment and rising trends from the call, turning hours of commentary into structured, granular insight
New trends, events and sentiment shifts
ProntoNLP's trend, event and sentiment tagging, compared across Micron's last five earnings calls.
- New trend: conventional DRAM. The term appears for the first time in the last five calls (4 mentions) and the trends engine flags it at 26x baseline. It tracks the HBM-versus-DDR debate: HBM prices catching up with conventional DRAM and analysts asking whether HBM wafers could shift back to DDR.
- New trend: generic memory talk is fading. NAND mentions are down 36% from a year ago (14 sentences vs 22) and DRAM down 52% (19 vs 40), even as revenue is up 6.3x and 4.4x. Product terms such as Data Center SSD are taking their place in the rising trends.
- Event shift: from revenue and contracts to pricing, capacity and capital. Versus June, pricing events rose from 6 to 11, shareholder return from 2 to 6 and capacity from 1 to 3, and CapEx events appeared (4). Revenue events fell from 32 to 17 and contracts from 8 to 4. CapEx drew 19 sentences, against 5 in June.
- Event shift: supply chain and market position turned positive. Supply chain events scored -0.36 in June (7 of 15 negative) and +0.56 now (2 of 16). Market position went from -0.38 to +0.91.
- Sentiment: tone is back, but the investment score is the swing. Sentiment is 0.83 (0.76 in June) and negative events are the fewest since September 2025 (25). The investment score recovered to 0.67 from 0.21 in June, still below 0.93 a year ago.
- Sentiment: caution sits in the Q&A and the financial outlook. Prepared remarks scored 0.93, management answers 0.72 and analyst questions 0.45, with 15 of 25 negative events in answers. Market forecasts scored 0.98 but financial forecasts only 0.31, where the negatives are cost commentary: about $1B of extra Q1 cost and Q1 as the gross margin floor.
- Sentiment: management is upbeat, analysts are split. The CEO scored 0.92 and the CFO 0.74, a reversal from June (0.68 and 0.91). Analysts averaged 0.20, with UBS at -0.80 (from +0.24) and TD Cowen at +0.80 (from -0.08).
- Sentiment vs. stock: tone has not predicted price. Next-day moves over the last five calls were -2.8%, +10.2%, -3.8%, +15.7% and +3.0%. June's low investment score came with the +15.7% jump, and the shares are down 14% since this call (to Oct. 14).
Source: ProntoNLP Transcript Analytics for Micron's last five earnings calls (Sep. 2025 to Sep. 2026); ProntoNLP price data. Scores range from -1 to 1; trend multiples are versus ProntoNLP's baseline; mention counts are sentences.
Sentiment Score Over Time
Alpha Signals Semiconductor Model
Powered by Alpha Signals Semiconductor Model
- Composite dropped after a record print. Semiconductor Model went from 9 on the call date to 14 by Oct 6 still ranking in the top quintile showing strong absolute results.
- Earnings momentum fell the most. EM Module went from 4 to 17, after holding at 1 through August. It ranks surprises and estimate revisions against peers, so the bar was high and revisions may lag the call.
- Management quality dipped on Oct 1. MQ Base Module went from 2 to 10. The 82.3% operating margin and $73.5B of cash support it. Inventory days of 129 and CapEx of about $25B in 1H FY2027 may weigh on working capital and cash burn.
- Demand Supply Ratio was flat. It stayed at 2 to 4 and was 3 on both Sep 30 and Oct 6. It measures stock borrowed against lendable inventory.
Semiconductor model ranks
Risks and headwinds
Management pointed to a few pressures on the outlook.
- Supply: New clean room capacity takes years to build, so tightness will persist.
- Cost pressure: Higher CapEx and incentive compensation weigh on near-term margins. Incentive compensation adds about $1B of cost in Q1.
- Inventory: Inventory days rose to 129, partly from build-ahead ahead of node end-of-life, but management expects them to fall.
- Customer behavior: Analysts asked whether big customers might cut memory content per server. Management said that is optimization, not weaker demand.
- Peak-cycle worries: Analysts asked whether depressed memory valuations signal a 2027 peak as capacity comes online. Management disagreed, pointing to tightness into 2028.
Rising Trends from ProntoNLP
Name |
Hits |
Baseline |
Score ▼ |
Change |
Data Center SSD |
3 |
1 |
1.00 |
↑ 12.1x |
NAND Flash Memory |
3 |
23 |
0.57 |
↓ -1.9x |
Clean Room Additions |
2 |
0 |
0.57 |
New |
HBM Products |
2 |
0 |
0.57 |
New |
AI_Workloads |
1 |
10 |
0.57 |
New |
G9 NAND Nodes |
1 |
0 |
0.57 |
New |
Workforce trends
Powered by Headcount Analytics
- The only organizational news was leadership additions, for high managerial positions aimed at faster, more integrated decision-making
- Overall tone was expansionary: higher incentive compensation for FY2027, increased CapEx, and new clean-room capacity buildout
- Headcount data corroborates this. Micron added headcount in every month of the trailing year (+4.7% overall), with growth accelerating from April through July 2026.
- Growth is concentrated in core manufacturing and R&D hubs (Boise headquarters, India, Singapore and Taiwan), consistent with capacity expansion. China is the notable area of weakness.
- Growth is broad across job levels. Senior roles added the most positions (+268, or +4.0%). Training (+77.6%) and Entry (+11.1%) grew fastest from small bases, and Owner level was the only notable decline (-7.1%).
Micron – Role Headcount and 12-month growth
About the datasets
ProntoNLP
ProntoNLP leverages advanced Natural Language Processing, using Large Language Models trained on S&P content to transform how users interact with market data. Clients can better understand market trends, access unique insights and data-driven recommendations, and get granular insight at the event and sentiment level without reading through the full document. Unlike general AI tools such as ChatGPT, which offer broad conversational capabilities, ProntoNLP is built specifically for market analysis and financial insights, delivering more accurate and relevant information tailored to industry needs.
Available as
Visible Alpha Estimates
The Visible Alpha Estimates dataset includes forecasts, assumptions, and logic from full working sell-side models. This deep consensus data provides a quick understanding of the sell-side view on a company or industry at an unprecedented level of granularity, timeliness, and interactivity.
Alpha Signals Semiconductors Model
The Alpha Signals Semiconductors Model is an operating-metric based factor model for the Semiconductor industry designed to generate additional alpha with greater consistency in positive return spreads. We built it upon our original Tech Model, it utilizes four semiconductor-specific indicators around Order Bookings & Order Backlogs in conjunction with a unique Short Sentiment indicator to capture the distinct characteristics of semiconductor firms.
Headcount Analytics
The Headcount Analytics dataset provides a comprehensive view of a company's workforce composition, trends, and metrics. It includes detailed insights into over 220 million employees worldwide, with monthly updates since 2010 covering more than 5.5 million entities. The dataset features geographic distribution of headcount data, employee tenure, and talent flow, enabling users to assess the operational structure of companies.