Fiscal Q4 blew through the guide and Q1 FY2027 guidance came in ~8% above consensus — but it was the raised capex bill, not the beat, that the tape traded — Quick Take
The screener route produced no new large-cap signal again. The HK VolumeSurge of Thu 1 Oct passed 2 of 2,817 names — and both are already in this series: 1093.HK CSPC Pharmaceutical (covered yesterday, 30 Sep) and 1787.HK Shandong Gold Mining (covered 21 Aug). The US VolumeSurge of Thu 1 Oct passed 31 names, dominated by Healthcare (13) and Technology (9) — but every passing name is sub-$15B. The three largest by turnover are GEN ($426M, ~$13B cap), ZETA ($352M) and FORM ($329M); the rest are small-caps and micro-caps. So the headline branch applied.
That headline is the largest single-name catalyst in today's feed. CNBC's lead corporate item for 30 Sep–1 Oct is "Micron beats on earnings and issues strong guidance as data center revenue jumps 11-fold" — the fiscal Q4 2026 print (quarter ended 3 Sep, reported after the close Wed 30 Sep). Revenue nearly quintupled, non-GAAP EPS of $33.42 beat the ~$31.61 consensus, and the Q1 FY2027 guide of ~$61.5B revenue / $38.15 EPS came in well above the $57.02B / $35.40 street estimate. Data-centre revenue jumped 11-fold. At ~$1.2T, Micron is the only US-based maker of HBM and among the largest S&P 500 constituents — the purest listed expression of the AI memory bottleneck.
Originally covered 28 Sep 2026 — three days ago — as an explicitly pre-earnings setup ("The AI Memory Supercycle Meets Its $50B Test"), which flagged the FQ4 print as a dated, binary catalyst. Re-researched under the protocol's earnings exception because that catalyst has now resolved: FQ4 revenue of $54.23B (+379% YoY), non-GAAP EPS $33.42 vs ~$31.61 expected, and a Q1 FY2027 guide ~8% above consensus. The prior take — "own it for the contracted cash flow, not for guidance day" — is now resolved in the affirmative on the fundamentals; what has changed the tape is the accompanying raised FY2027 capex plan, which overshadowed the beat after hours.
| Year | Revenue | Rev Growth | Op Margin | Net Income |
|---|---|---|---|---|
| FY2022 | $30,758M | +11% | 31.6% | +$8,687M |
| FY2023 | $15,540M | -50% | -23.0% | -$5,833M |
| FY2024 | $25,111M | +62% | 5.2% | +$778M |
| FY2025 | $37,378M | +49% | 26.1% | +$8,539M |
| FY2026 | $133,188M | +256% | 74.6% | +$84,969M |
Fiscal years end in early September. The shape of this table is the memory cycle: FY2023 halved revenue to $15.5B with a $5.8B net loss, and three years later FY2026 revenue is $133.19B with GAAP net income of $84.97B ($74.33 per diluted share). The step-change is entirely in the last twelve months. FQ4 2026 delivered revenue of $54,229M (versus $41,456M in FQ3 and $11,315M a year earlier), GAAP gross margin of 86.8%, operating income of $43.75B (80.7% of revenue), and net income of $37.70B ($32.87/sh GAAP, $33.42/sh non-GAAP). Operating cash flow was $43.97B against $10.77B of net capex, for $33.20B of adjusted free cash flow in a single quarter.
The mix is the story: Core Data Center revenue of $18.00B at a 90% gross margin (85% operating margin) and Cloud Memory of $16.28B at 83% GM (76% OM) are now more than 60% of the company, with Mobile & Client at $13.11B and Automotive & Embedded at $6.82B. DRAM revenue specifically rose 343% YoY to $39.8B — 73% of total sales. For the year, operating cash flow was $89.68B, capex $27.37B, adjusted free cash flow $62.31B, and the company ended with $73.48B of cash, marketable investments and restricted cash — a net-cash fortress versus the $3.4B of net debt at FY2025 year-end.
Guidance is where the argument now lives. Q1 FY2027 is guided to $61.5B ± $1.5B of revenue (≈$57.02B consensus), ~86% gross margin, and non-GAAP EPS of $38.15 ± $1.00 (≈$35.40 consensus) — an ~8% EPS beat at the mid-point. But management simultaneously raised FY2027 capex (≈$11.5B in FQ1, ≈$25B in H1 FY27, versus $27.37B for the whole of FY2026) to expand DRAM cleanroom capacity, on top of a $250B multi-year program building two new HBM campuses (Clay, NY broke ground in January; the first new Boise fab comes online next year). Management expects memory and storage supply-demand to be "much tighter" in FY2027 and FY2028, with capital return stepped up primarily through buybacks. The tape's read: after a >500% twelve-month run, the shares closed at $1,074.95, were fractionally lower after hours ($1,071.92, -0.28%) and have traded roughly $1,041–$1,089 on 1 Oct — a huge beat that barely moved the stock.
Valuation at ~$1,075 is roughly 14.5x trailing GAAP EPS ($74.33) but only ~7x annualised forward earnings (FQ1 guide ×4), ~9x FY2026 sales and ~19x FY2026 free cash flow. The dividend was held at $0.15/quarter ($0.60 annualised), declared 30 Sep. Morgan Stanley kept Overweight with a $1,200 target; Mizuho reiterated Buy; the CNBC Investing Club raised its target into the print.
Micron just printed the cleanest quarter in its history — $54.2B of revenue, an 80.7% operating margin, $33.2B of quarterly free cash flow — and guided Q1 FY2027 ~8% above consensus, which makes ~7x annualised forward earnings genuinely hard to argue against. The risk is not the quarter; it is that the entire industry is now funding a supply response on the same demand signal, and after a >500% year the stock's refusal to move on a blowout is the market quietly saying it already believes the cycle is closer to the peak than the trough. Own it for contracted HBM demand and the FCF yield, treat the raised capex number as the single most important line in the release — and watch DRAM pricing, not the headline beat, for the first sign that this one is rolling over.
Disclaimer: AI-generated quick take for informational purposes only. Not investment advice. Data from public filings and financial data providers.