MU
Micron Technology, Inc.
HBM Memory · AI Chips & Compute
📌 Current holding · 8.00% NLV
~$777
Mkt Cap ~$877B
52W: $83.37 – $818.67
Next earnings: ~24 Jun 2026 (Q3 FY26)
⏸ Hold · Do Not Add Here
Add zone: $600–650 · Trim trigger: $850+
⚠ EXTENDED PRICE ALERT: +68% YTD, sitting at 52W highs ($818 intraday 12 May). Consensus PT $483–533 now well below spot. Position is 8.00% NLV — correctly sized. Do not add. Define a trim trigger at $850–900 before the 24 Jun print.
The market is still pricing MU on commodity memory cycle logic — but the setup is structurally different this cycle: HBM3e and HBM4 supply through end-2026 is fully contracted under take-or-pay agreements, NAND heavyweight customers are securing supply-and-contract agreements to stabilise pricing, and management's Q3 single-quarter revenue guidance already exceeds Micron's full-year revenue for every fiscal year through 2024. The secular re-rating thesis — memory becoming less cyclical as AI infrastructure demand creates multi-year contracted floors — is now backed by operating results, not just narrative. The unresolved question is whether 2027–2028 fab capacity convergence re-commoditises the market before the product mix shift to HBM4/HBM4E compounds.
Fwd P/E (NTM)
7.7–8.1×5
PEG <0.2 (g=196% NTM) ⚠ cycle-peak g
Rev YoY (latest Q)
+196.3%8
QoQ +74.9% · $23.9B
Gross Margin
74.9%8
adj · 74.4% GAAP · Q3g 81% ↑
FCF Margin (Q2)
~29%8
adj FCF $6.9B · OCF $11.9B (49.9%)
EPS Revision (90d)
↑ strong8
Q3g $19.15 beat est +77.8% · ⚠ analyst count tbc
Next Earnings
24 Jun
Q3 FY26 · Confirmed2
AI DC Rev Mix
56%8
CMBU $7.75B · CDBU $5.69B
DC Rev YoY
$33.5B Q3g2
~200% YoY · ±$750M
Thesis Durability
A
Variant perception remains intact: market still applies commodity-cycle discount to MU while HBM supply is locked under multi-year take-or-pay contracts. Three-player oligopoly (SK Hynix, Micron, Samsung) with no fourth entrant possible before 2029 due to fab lead times. Sole US-based HBM supplier — CHIPS Act structural advantage. CEO confirmed 100% of calendar 2026 HBM supply (including HBM4) is under binding pricing/volume agreements:1
NAND is no longer an afterthought: historic pricing surges are pushing enterprise customers to seek stability via supply-and-contract agreements (SCAs) — the same contracted-floor dynamic now emerging in NAND that HBM pioneered. Longer context windows, reasoning, and agentic workloads are also sustaining DRAM demand beyond the AI-training cohort.8
Management on the Q2 FY26 print:
"We have completed agreements on price and volume for our entire calendar 2026 HBM supply, including Micron's industry-leading HBM4." — Sanjay Mehrotra (CEO), Q1 FY26 Earnings Call, 18 Dec 2025TAM: HBM market growing from $35B (2025) → $100B (2028) at 40% CAGR — accelerated from prior $80B-by-2030.2 AI accelerators (NVDA Blackwell/Rubin, AMD MI450) consuming more HBM per chip generation.
NAND is no longer an afterthought: historic pricing surges are pushing enterprise customers to seek stability via supply-and-contract agreements (SCAs) — the same contracted-floor dynamic now emerging in NAND that HBM pioneered. Longer context windows, reasoning, and agentic workloads are also sustaining DRAM demand beyond the AI-training cohort.8
Management on the Q2 FY26 print:
"Quarterly revenue nearly tripled versus one year ago, and revenue for DRAM, NAND, HBM and each business unit reached new highs. Our fiscal Q3 single quarter revenue guidance exceeds the full year revenue for every year in our company's history through fiscal 2024. For fiscal Q3, we anticipate exceptional records across revenue, gross margin, EPS and free cash flow." — Management, Q2 FY26 Earnings Call (as cited in IO Fund Q2 2026 Top 15)8Bear case: Samsung resolves HBM3e yield issues by H2 2026, re-entering as a credible 3rd supplier. New fab capacity from all three players enters late 2027. Survivable — contracts extend visibility, but 2028 is a legitimate re-pricing risk.
Thesis is evolving from narrative to operating proof. The contracted-floor argument — once a forward-looking call — is now backed by a Q2 print that pulled forward the $33B quarterly revenue milestone by six quarters vs prior roadmap estimates. The incremental thesis extension: NAND secular demand (SCAs) broadens the contracted-floor thesis beyond HBM.
Business Quality
A
Q2 FY26: Revenue $23.9B (+196.3% YoY, +74.9% QoQ) — largest sequential dollar increase in the company's history at +$10.2B.8 GAAP EPS $12.07 beating estimates by 36.3%; adj EPS $12.20 beating estimates by 36%.8 GAAP net income $13.79B. OCF $11.9B (49.9% margin); adj FCF $6.9B (28.9% margin). Adj gross margin 74.9% — company record, beating prior guidance of 67% by 790bps. Q3 guidance: $33.5B revenue (±$750M), 81% adj gross margin, adj EPS $19.15 — all-time records on every line.8
AI data-centre breakdown: Cloud Memory Business Unit (CMBU) $7.75B (+163% YoY, +47% QoQ); Core Data Centre Business Unit (CDBU) $5.69B (+211% YoY, +139% QoQ). Combined AI DC revenue $13.44B = 56% of Q2 total. CMBU+CDBU Q3 guide implies ~40% QoQ on similar mix.8
Capex: Q2 actual $5.0B (+61.3% YoY); Q3 guide $7.0B — capex intensity rising with HBM4 ramp and Idaho/New York fabs. Adj FCF expected to roughly double sequentially in Q3 vs Q2.8 Sell-side estimate revisions sharply up post-print.
AI data-centre breakdown: Cloud Memory Business Unit (CMBU) $7.75B (+163% YoY, +47% QoQ); Core Data Centre Business Unit (CDBU) $5.69B (+211% YoY, +139% QoQ). Combined AI DC revenue $13.44B = 56% of Q2 total. CMBU+CDBU Q3 guide implies ~40% QoQ on similar mix.8
Capex: Q2 actual $5.0B (+61.3% YoY); Q3 guide $7.0B — capex intensity rising with HBM4 ramp and Idaho/New York fabs. Adj FCF expected to roughly double sequentially in Q3 vs Q2.8 Sell-side estimate revisions sharply up post-print.
Insider flag:
- CEO Sanjay Mehrotra sold 40,000 shares @ $536 avg on 2026-05-01 (~$21.5M, -8.6% of direct stake to 424,503).4 Pre-planned 10b5-1 adopted 2026-01-30 — not a discretionary bearish signal.
- 18-month net pattern: 984,581 shares sold vs 23,200 bought across all insiders. Consistent with monetising appreciated equity, not thesis change.
- No single sale >$50M ex-Mehrotra; no cluster of multiple insiders moving same direction in close succession.
Pristine operational execution. Record margins, record FCF, upward estimate revisions. Insider selling is plan-driven, not directional.
Entry Price Discipline
C+
At $777, MU trades at:
Institutional positioning: 26/27 sell-side at Buy, but consensus PT lagging by 35%. Not a sell signal — means upside is now dependent on 24 Jun print beating guidance convincingly, not on consensus reprice.7
- Fwd P/E ~7.7–8.1× NTM (FinanceCharts 7.72×)5 — cheap in absolute terms vs sector median ~35×
- FY2026 adj EPS estimate $57.8 (+597% YoY) → FY26 P/E ~13.4× at $777 · FY2027 NTM higher still8
- TTM P/E ~20× — at high end of 3Y avg 18.85× and 5Y avg 19.53×5
- +68% YTD; 52W intraday high $818 (12 May)
Institutional positioning: 26/27 sell-side at Buy, but consensus PT lagging by 35%. Not a sell signal — means upside is now dependent on 24 Jun print beating guidance convincingly, not on consensus reprice.7
Not an add at $777. Risk-reward asymmetric only if Q3 beats with Q4 guide >$36–38B. Momentum stock characteristics have emerged — this is now a hold-discipline test.
Verdict — Existing Holding
Hold at current 8.00% NLV. The original entry thesis is intact and well in the money — but the stock has moved from the add zone ($600–650) to $777 in a parabolic run, and the 24 Jun Q3 print is now fully the next decision point. A Q3 beat (revenue >$35B, Q4 guide >$36B, constructive HBM4 qualification language) validates the next leg. A miss or conservative Q4 guide creates significant mean-reversion risk from extended momentum highs.
The right discipline is to do nothing now, define a trim rule, and decide at the June print: if Q3 beats with conviction → hold. If Q3 is in-line or guide disappoints → trim ~30% of the position at the post-print price.
Cyclicality reminder: every Hold/Add on MU must flag that 75–81% gross margins are not the steady-state. The downside in a Samsung-resolves-yields + 2027-capacity scenario is a 30 ppt margin compression — modelled in the DCF bear at $280–320 fair value.
The right discipline is to do nothing now, define a trim rule, and decide at the June print: if Q3 beats with conviction → hold. If Q3 is in-line or guide disappoints → trim ~30% of the position at the post-print price.
Cyclicality reminder: every Hold/Add on MU must flag that 75–81% gross margins are not the steady-state. The downside in a Samsung-resolves-yields + 2027-capacity scenario is a 30 ppt margin compression — modelled in the DCF bear at $280–320 fair value.
Current 8.00% NLV — hold. Hard rule: do not add above $760. Trim trigger: trim ~25% of the position if stock exceeds $870 before 24 Jun print (locks gains on roughly a quarter of the holding). Post-24 Jun: reassess full sizing. 20% single-stock cap is not the binding constraint here — price discipline is.
Opportunity-cost check
▾
vs CSPX (40% NLV)
At 8× forward P/E with ~200% quarterly revenue growth, MU at hold level still has structural upside over a passive index position — HBM contract lock-in provides revenue visibility CSPX exposure cannot replicate. Edge is AI memory oligopoly pricing power, not momentum.
vs current holdings
MU sits alongside PLTR, NOW, META, GOOGL. Provides distinct AI infrastructure layer exposure (memory) not covered by software names. Adding more MU at $777 vs deploying cash into NOW (5.53%, below target) is the cleaner opportunity-cost trade.
Thesis-breakers
▾
Three falsifiers. Any one trips → trim ~50% of position (target ~4% NLV). Any two → exit to 3% NLV.
- 01HBM ASP decline — if 24 Jun commentary signals HBM3e/HBM4 pricing pressure >10% QoQ, the locked-contract thesis has broken · By Q3 FY26 print (24 Jun 2026)
- 02Q3 FY26 revenue misses $33.0B (below guide midpoint by >$500M) — indicating demand pullback or supply execution miss; collapses bull consensus and triggers 20–30% drawdown · By Q3 FY26 print (24 Jun 2026)
- 03Any hyperscaler guides AI capex flat or down for 2027 (MSFT, GOOGL, META, AMZN) — HBM demand is 100% a function of AI infrastructure spend; capex pause reprices 2027–2028 memory demand · By Sep 2026
Key risks
▾
- 01Re-commoditisation 2027–2028: Samsung HBM yield fix + Micron Idaho fab + SK Hynix expansion all converge late 2027. Three fully-ramped HBM suppliers chasing plateauing demand reverts ASPs toward commodity DRAM margins (30–40% vs current 75–81%). A 30 ppt GM compression on $130B run-rate is a ~$40B gross profit hit. MU would derate toward 5–7× trailing P/E.
- 02Momentum unwind risk: Stock +50% in 2 weeks, at 52W highs, RSI overbought. Parabolic move from ~$480 (early May) to $818 intraday has outpaced any fundamental repricing. Reversion to $600–650 on profit-taking or slight disappointment = 16–23% drawdown.
- 03Samsung HBM yield normalisation: Samsung was the "weak link" that kept the 3-player oligopoly effectively 2-player. If Samsung resolves HBM3e qualification with NVIDIA in H2 2026 (flagged as likely by SemiAnalysis), supply tightness eases ahead of 2027–2028 new fab capacity — compressing the window of oligopoly pricing power.
- 04Consensus PT lag creates false ceiling: Street PT ~$483–533 while stock trades $777. Either analysts haven't updated models for the parabolic run, or analysts believe a 35–40% drawdown to fair value is coming. Rely on 24 Jun print fundamentals, not stale PTs.
- 05Capex execution risk: Q3 guide $7.0B capex (Idaho + New York fabs). CHIPS Act partially offsets, but execution delays, HBM4 yield ramp issues, or macro tightening raising debt costs could compress FCF even as revenues grow.
- 06Vera Rubin air pocket: Micron's HBM and data-centre DRAM sit earlier in the AI supply chain than Nvidia's peak GPU shipment revenue. Even with the secular trend intact, demand gaps tied to Vera Rubin production ramp cadence can create quarterly revenue air pockets — particularly if GPU shipment volumes cluster unevenly. The secular thesis and a near-term air pocket can both be true simultaneously.8
DCF scenarios
▾
Discount rate: 12% — high-growth profitable semiconductor; higher than NVDA (10–11%) due to memory cyclicality and capex intensity.
FCF base (FY26e)
~$40B
5Y FCF CAGR
25%
Terminal growth
4%
Fair value / share
~$980
Bull: HBM pricing holds through 2027 + Samsung stays excluded from NVDA Rubin supply chain. Adj FCF: Q2 $6.9B, Q3e ~$13.8B (expected to roughly double), Q4e ~$14B → FY26 adj FCF ~$35–40B. Ramps to $100B+ by FY30 as revenues scale. IO Fund FY26 adj EPS estimate $57.8 (FY27 likely higher given Q3 margins). D.A. Davidson $1,000 PT aligns. Implies +26% upside from $777.8
FCF base (FY26e)
~$40B
5Y FCF CAGR
−10% CAGR
Terminal growth
2%
Fair value / share
~$280–320
Bear: Memory cycle reverts in 2027 as Samsung + new capacity re-enters. Gross margins compress 81% → 45% by FY28. FCF declines 40–50% from peak. Simply Wall St DCF implies $268 fair value; GF Value bear ~$335. Implies −60% downside — the memory cycle playbook. Key assumption that makes this wrong: HBM locked contracts hold.
Position: At $777, stock sits between the two scenarios — above the bear ($280–320) and below the bull ($980). Thesis hold = current price fair-to-cheap. Bear materialising = 60% drawdown from here.
Catalyst timeline
▾
-
12 MayMU at JEDEC forum — mobile/edge track. Early read on HBM4 qualification status with NVIDIA Rubin architecture.
-
13 MayMU at JEDEC forum — server/cloud AI track. Follows mobile/edge session; combined read on HBM supply-chain positioning.
-
20 MayNVDA Q1 FY27 earnings — primary read on Blackwell demand and HBM consumption per GPU. NVDA beat + strong Rubin guide = bullish MU. Miss or capex guide-down = immediate MU pressure.
-
24 JunMU Q3 FY26 earnings (guide: $33.5B ±$750M rev, ~81% GM, EPS $19.15). Key questions: (1) Does HBM pricing commentary confirm 2027 stability? (2) Is Q4 guide >$36B? (3) HBM4 certification timeline with NVIDIA. This print decides the next sizing call.
-
2026-Q3Samsung HBM3e qualification update — if Samsung is cleared for NVIDIA Rubin supply, the effective 2-player oligopoly becomes 3-player, and 2027 pricing assumptions need re-modelling.
Named analyst commentary
▾
"I foresee Micron becoming more secular than the market has historically treated it."
— Beth Kindig / IO Fund — Q2 2026 Top 15, "Micron: Doors. Blown. Off." (io-fund.com, paywalled · May 2026)8
"The difference between the AI cycle and the cyclical peaks in the past is that Micron is combining record fundamentals with improving visibility through multi-year customer agreements and a strengthening product roadmap (HBM4/HBM4E, 1γ DRAM, Gen6 SSDs)."
— Beth Kindig / IO Fund — Q2 2026 Top 15 (May 2026)8
"As someone who looks at hundreds of earnings reports a year... I cannot recollect seeing one quite like this. Micron blew the doors off with revenue growth of 196.3% YoY and up 75% QoQ for a beat of 22.3% on a massive revenue base of about $24 billion a quarter."
— Beth Kindig / IO Fund — Q2 2026 Top 15 (May 2026)8
No commentary from Serenity (@aleabitoreddit) or Dylan Patel / SemiAnalysis on MU located as of 13 May 2026.
Sources
1 Micron IR — Q1 FY26 Earnings Call transcript, December 2025
2 Micron 8-K, Q2 FY26, filed 2026-03-18 — SEC EDGAR. Revenue $23.86B, EPS $12.20, Q3 guide $33.5B / 81% GM / $19.15 EPS
3 StockAnalysis.com / TradingView — consensus beat data, Q2 FY26 ($12.20 vs $9.19 est, 32.8% beat)
4 SEC Form 4 — Sanjay Mehrotra, filed 2026-05-05. 40,000 shares sold at $536 avg under 10b5-1 plan adopted 2026-01-30
5 FinanceCharts.com — historical P/E data. Fwd P/E 7.72 as of May 2026; 3Y avg 18.85; 5Y avg 19.53
6 TechMarketer / Watcher.guru — D.A. Davidson $1,000 PT (Street-high); consensus range $478–660 (33 analysts)
7 24/7 Wall St — Wall Street consensus PT ~$533 as of May 2026; 26/27 analysts at Buy
8 Beth Kindig / IO Fund — Q2 2026 Top 15 report, "Micron: Doors. Blown. Off." (io-fund.com, paywalled); captured 2026-05-13. Where IO Fund figures differ from secondary aggregators, IO Fund figures are used as the trusted primary source for reported/guided fundamentals.
⚠ Forward P/E varies across providers (GuruFocus 8.05×, FinanceCharts 7.72×, Macrotrends ~20× TTM) due to differing NTM vs TTM definitions. Range 7.7–8.1× used for NTM Fwd P/E; FY2026 full-year adj P/E ~13.4× at $777 (using IO Fund FY26e adj EPS $57.8).
1 Micron IR — Q1 FY26 Earnings Call transcript, December 2025
2 Micron 8-K, Q2 FY26, filed 2026-03-18 — SEC EDGAR. Revenue $23.86B, EPS $12.20, Q3 guide $33.5B / 81% GM / $19.15 EPS
3 StockAnalysis.com / TradingView — consensus beat data, Q2 FY26 ($12.20 vs $9.19 est, 32.8% beat)
4 SEC Form 4 — Sanjay Mehrotra, filed 2026-05-05. 40,000 shares sold at $536 avg under 10b5-1 plan adopted 2026-01-30
5 FinanceCharts.com — historical P/E data. Fwd P/E 7.72 as of May 2026; 3Y avg 18.85; 5Y avg 19.53
6 TechMarketer / Watcher.guru — D.A. Davidson $1,000 PT (Street-high); consensus range $478–660 (33 analysts)
7 24/7 Wall St — Wall Street consensus PT ~$533 as of May 2026; 26/27 analysts at Buy
8 Beth Kindig / IO Fund — Q2 2026 Top 15 report, "Micron: Doors. Blown. Off." (io-fund.com, paywalled); captured 2026-05-13. Where IO Fund figures differ from secondary aggregators, IO Fund figures are used as the trusted primary source for reported/guided fundamentals.
⚠ Forward P/E varies across providers (GuruFocus 8.05×, FinanceCharts 7.72×, Macrotrends ~20× TTM) due to differing NTM vs TTM definitions. Range 7.7–8.1× used for NTM Fwd P/E; FY2026 full-year adj P/E ~13.4× at $777 (using IO Fund FY26e adj EPS $57.8).