Theta Research
  • Home
  • Graph
  • CompaniesPro
  • Pulse
  • Pricing
  • Shipped
  • Engine

For serious investors who think in systems.

partners@theta.md

Xiaohongshu · 26469493287

ProductAtlasCompaniesPulseEventsEnginePricing
CompanyAboutHow it worksChangelog
LegalTermsPrivacyRefunds

© 2026 theta.md

Theta Research provides research and analytics for informational purposes only. Nothing on this site is investment advice.

← Back to Events
COMPLETEDTier 1earnings-single

Micron Q3 FY2026 Earnings — Beat-vs-Elevated-Consensus + HBM4 Allocation Quantification (the sell-the-news test at an all-time high)

2026-06-24 → 2026-06-24·11 companies·12 predictions·Brier: 0.193
—MUdirect?SKHynixcompetitive?Samsungcompetitive?SNDKcompetitive?WDCcompetitive?NVDAsupply-chain?AVGOsupply-chain?LRCXsupply-chain?AMATsupply-chain?ASMLsupply-chain?KLACsupply-chain

Scenario Comparison

Bull CaseS1
30%

Q3 revenue clearly beats the elevated Street consensus (≥ around $37B vs $35.88B consensus / $33.5B guide), non-GAAP GM ≥81% (meets/exceeds the unprecedented guide), non-GAAP EPS beats consensus (≥ around $21 vs $20.69). Q4 FY2026 guide STEPS UP to roughly $38B+ with GM held around 80%+, AND management QUANTIFIES a materially larger/growing Vera Rubin HBM4 share or FY2027 HBM pre-booking. The bull-confirming, pop-and-hold outcome.

↑ 9↓ 011 companies
+MUstrong
+SKHynixstrong
+Samsungmoderate
~NVDAmild
+7 more
Base CaseS2
47%

Q3 beats the guide and roughly meets-to-modestly-beats the elevated Street consensus (revenue around $35-36.5B, EPS around $20.5-21, GM 80-81%) BUT the Q4 FY2026 guide is only IN-LINE to modestly higher (no clear step-up to around $38B) and management does NOT quantify a larger Vera Rubin HBM4 share (reiterates residual #3 position vaguely). HBM run-rate reaffirmed but not raised.

↑ 0↓ 110 companies
−MUmoderate
~SKHynixmild
~Samsungmild
~NVDAmild
+6 more
Bear CaseS4
3%

Q3 revenue misses the GUIDE itself (<$32.8B, >2% below the $33.5B guide) OR non-GAAP GM <78% (>300bp below the 81% guide), AND Q4 FY2026 guidance is cut materially below the Q3 run-rate. Management discloses a concrete demand/pricing shock (hyperscaler order push-out, aggressive Samsung HBM4 undercut, or unexpected inventory build).

↑ 0↓ 910 companies
−MUstrong
−SKHynixstrong
−Samsungstrong
−LRCXstrong
+6 more

All Scenarios

6

Positioning Suggestions

→ HOUSE STANCE = HOLD. Do NOT chase MU at $1,192.84 (June 22, all-time high, +5.19% today). The risk/reward into the print is unfavorable: the price is already ABOVE the $1,150 12-month base target; Street consensus ($35.88B / $20.69) sits above Micron's own guide ($33.5B / $19.15), so the market is pricing a beat-vs-elevated-consensus; and sentiment is exhausted (18 strong-buy / 33 buy / 3 hold / 1 sell). Only the around 30% Scenario-1 outcome (blowout beat + strong Q4 guide step-up + HBM4 share QUANTIFIED) justifies chasing.

→ Split the FUNDAMENTAL print from the STOCK reaction. Fundamentals: the beat is high-confidence (4-quarter beat streak, ASP inflation, sold-out HBM, supply crunch). Stock reaction: sell-the-news SKEWED because consensus is above guide, recs are exhausted, the stock is already +5% today at an all-time high, and much is priced at peak-cycle P/B (around 12x). A fade on a beat (Scenario 2) is the single most-likely outcome (around 47%).

→ UPGRADE path (thesis gate): move toward Buy on a pullback to roughly $850-900 (where the bull/base asymmetry turns clearly favorable) OR on a Q3 print that quantifies a materially larger/growing Vera Rubin HBM4 share. Until one of those fires, do not add at the all-time high.

→ DOWNGRADE path (thesis gate): move toward reduce if Q4 FY2026 guide comes in below around $32B, GM rolls below around 55% for two quarters, or DDR5/CXMT oversupply signals appear earlier than 2027. Scenario 3-4 (soft guide / GM rollover / share-reallocation wobble / miss) fires these — expect a multi-quarter de-rate toward the 12-month base $1,150 then bear $620.

→ Vera Rubin HBM4 share QUANTIFICATION (Scenario 5) is the swing disclosure: a number showing a materially larger/growing MU share is the upgrade trigger and can rescue a sell-the-news tape; continued vagueness maintains the #3 discount and keeps the stance HOLD.

→ Pair/read-through: WFE names (LRCX highest beta, then AMAT) are leading indicators — the current 'supply crunch' narrative (Citi: AMAT $710, LRCX $450) confirms tight supply, but a soft MU capex/guide read-through would de-rate them faster than MU itself. SNDK is the highest-beta NAND read-through (up on Scenario 1, down on Scenario 3-4).

→ Avoid catching a knife on a Scenario-3/4 overshoot: a soft guide or GM rollover off an 81% GM at an all-time-high multiple can gap MU sharply; let the de-rate find a base near the $1,150 base / $850-900 add-zone before re-evaluating a long.

→ Memory peers (SKHynix/Samsung) are thinly-covered read-through proxies here — use them to confirm the sector signal, not as primary positions. SK Hynix is now South Korea's most valuable company; the sector is euphoric.

Predictions

12
✓

MU Q3 FY2026 revenue prints ≥ the official $33.5B guide (record source: Micron IR) — i.e., at/above the company's own guidance.

92%P1

If revenue < $33.5B (a guide miss off a 4-quarter beat streak) → cycle-roll signal at an all-time-high multiple; expect a sharp gap-down toward the $1,150 base. Very low probability.

✓

MU Q3 FY2026 revenue beats the elevated Street consensus of $35.88B (third-party, Finnhub aggregator) — i.e., clears the bar the market is actually pricing, not just the guide.

52%P2

Deliberately a coin-flip: consensus sits +7% above Micron's own guide, so a beat-vs-consensus is far from assured even with a strong print. Failing to clear consensus is the core sell-the-news mechanism.

✓

MU Q3 FY2026 non-GAAP diluted EPS beats the Street consensus of $20.69 (third-party, Finnhub) — vs the official $19.15 guide (record source).

60%P3
✓

MU Q3 FY2026 non-GAAP gross margin prints ≥ 81% (meets/exceeds the unprecedented official guide).

62%P4

If < around 78% → peak-margin question opens immediately; a GM rolling below around 55% for two quarters is the thesis downgrade gate. Watch the Q4 GM guide more than the Q3 print.

✓

MU Q4 FY2026 revenue GUIDANCE midpoint ACCELERATES to ≥ roughly $38B (a clear sequential step-up above the Q3 run-rate).

40%P5

THE BULL CONFIRMATION. A ≥ around $38B Q4 guide (Scenario 1) is the step-up that justifies chasing; an in-line/modest guide is the sell-the-news fade (Scenario 2, most likely). A guide < around $32B is the thesis DOWNGRADE gate (Scenario 3).

✓

MU Q4 FY2026 non-GAAP gross margin guidance midpoint is ≥ 79% (peak GM sustained rather than normalizing toward high-70s/lower).

55%P6

If guided below around 78% → peak-GM normalization; combined with a GM trajectory below around 55% over two quarters this fires the thesis downgrade gate toward reduce.

✓

Management QUANTIFIES a materially larger or growing Micron Vera Rubin VR200/NVL72 HBM4 share / FY2027 HBM pre-booking on this call (a number, not just 'certified supplier' framing).

35%P7

THE THESIS UPGRADE TRIGGER. If MU quantifies a materially larger/growing share → move toward Buy (it lifts the FY2027 HBM trajectory toward the $1,500 bull). Continued vagueness (the more likely path) maintains the #3 discount and keeps the stance HOLD. Note: the prior 'lockout / MU = 0%' premise is RETIRED — MU is a certified supplier per NVIDIA (Jensen Huang, June 5 2026).

✓

Management raises or reaffirms (does not cut) the HBM annualized revenue run-rate.

78%P8
○

Management confirms FY2027 capex stepping up toward approximately $35B (within the multi-year around $200B cycle, above FY2026's >$25B).

55%P9

A confirmed around $35B reads bullish near-term (demand-responsive) but IS the operating-leverage exposure if FY2027 oversupplies (CXMT + new-fab output) — thesis-central, not thesis-breaking.

✓

Management's FY2027 supply-demand framing remains reassuring (shortage/tightness extending into 2027-2028) rather than introducing explicit CXMT / three-vendor oversupply caveats.

60%P10

If oversupply caveats appear EARLIER than 2027 → fires the thesis DOWNGRADE gate (the FY2027 oversupply seed pulled forward); move toward reduce.

○

MU stock closes June 25, 2026 (T+1 to the print) ABOVE its June 24 pre-print close (i.e., the beat pops and holds rather than fading).

47%P11

Deliberately a coin-flip given the run-up: the stock is already +5% today at an all-time high (around $1,193), above the $1,150 base, with consensus above guide and recs exhausted. A sell-the-news fade even on a beat is the single most-likely outcome (Scenario 2).

○

MU stock closes June 25, 2026 (T+1) DOWN more than 8% vs the June 24 pre-print close (sharp sell-the-news or guide-down reaction).

20%P12

If triggered alongside a soft Q4 guide / GM rollover (Scenario 3-4) → the downgrade-gate drawdown is underway; expect a multi-quarter de-rate toward the $1,150 base / $850-900 add-zone.

Key Questions

  1. BEAT-VS-WHAT: Q3 revenue/EPS clear the GUIDE (record source $33.5B / $19.15) with high confidence — but do they clear the ELEVATED Street consensus (Finnhub $35.88B / $20.69, which sits +7%/+8% ABOVE the guide)? The bar is the consensus, not the guide; that gap is the core sell-the-news risk.
  2. Q4 FY2026 revenue GUIDE direction: does it step UP toward roughly $38B+ (bull-confirming), come in IN-LINE/modest (sell-the-news fade risk even on a beat), or roll DOWN below around $32B (thesis downgrade gate)?
  3. 81% non-GAAP GM: does it print, and — more important — is Q4 GM guided to HOLD around 80%+ (peak sustained) or normalize to high-70s/lower? GM rolling below around 55% for two quarters is a thesis downgrade gate.
  4. VERA RUBIN HBM4 SHARE QUANTIFICATION (the lockout is retired): MU is a certified supplier — does management put a NUMBER on its Vera Rubin HBM4 share / FY2027 HBM pre-booking? A materially larger/growing share is the thesis UPGRADE trigger; continued vagueness maintains the #3 discount.
  5. HBM revenue run-rate + mix: is the HBM run-rate raised, and what % of DRAM revenue is HBM? Three-vendor ASP durability is the swing on FY2027 HBM dollars.
  6. DRAM bit-shipment growth + blended ASP QoQ: still accelerating (Apple-Cook 'unavoidable' memory price hikes, record DRAM/NAND ASPs), or first signs of deceleration?
  7. FY2027 capex: is the step toward around $35B confirmed within the multi-year around $200B cycle? Bullish demand signal near-term, operating-leverage exposure if FY2027 oversupplies (CXMT +120-140K wpm China DRAM reaching yield).
  8. FY2027 supply-demand framing: does management reassure (shortage through 2028) or introduce CXMT / three-vendor-capacity caveats? An earlier-than-2027 oversupply signal is a thesis downgrade trigger.
  9. Micron-Anthropic partnership (announced June 22): any color on scope/scale of the AI-infrastructure tie-up that drove the +5.5% pre-print pop — incremental demand, or sentiment-only?
  10. NAND commentary + read-through: does MU confirm NAND ASP strength (broad memory rally, WDC +6% / SNDK +5% last 7d), and what's the inventory read-through to SNDK/WDC?
  11. Tariff impact: how much margin pressure from the 10-50% module/SSD duties (bare chips exempt), and is customer absorption holding?

Monitoring Checklist

0/14 checked

Post-Event Results

Brier Score0.193
correct9/9
Trade Actions1

Actual Outcomes

  • • Q3 FY2026 revenue $41.46B — +346% YoY, +74% QoQ, ~+24% above the $33.5B guide and ~+15% above the $35.88B Street consensus. A blowout exceeding even pre-event Scenario 1's ≥$37B threshold.
  • • Non-GAAP gross margin 84.9% (GAAP 84.6%) vs the unprecedented 81% guide — beat. Non-GAAP EPS $25.11 (GAAP $24.67) vs $20.69 consensus / $19.15 guide.
  • • Q4 FY2026 guidance: revenue $50.0B (+20% QoQ), GM ~86%, non-GAAP EPS $31.00 — a massive sequential step-up, far above Scenario 1's $38B+ bull threshold. The single biggest surprise.
  • • Strategic Customer Agreements quantified: 16 multi-year SCAs, ~$22B customer cash deposits, ~$100B RPO — a structural forward-booking disclosure that reframes the memory revenue stream toward contract-visibility. Commitment to return 100% of excess cash.
  • • HBM: 2026 sold out, contracts into 2027; HBM4 12-high ramping ~2x HBM3E pace, >$1B HBM4 revenue, in HVM to lead customer (NVDA Vera Rubin).
  • • Regular-session reaction muted (close $1,048.51, -0.37%, pre-print) because earnings hit after close; after-hours ~+13% (~$1,180). T+1 (June 25) close is the definitive sell-the-news vs pop-and-hold test — still pending.

Market Reaction

MU-0.4%

Trade Recommendations

→ Summary: Scenario 1 (blowout beat + strong Q4 step-up + HBM/SCA quantification, pre-event prob 0.30) is what happened — and the actuals exceeded even its thresholds. 9/12 predictions resolved, all TRUE (avg Brier 0.193); 3 unresolved (FY2027 capex figure not disclosed; two T+1 stock-reaction predictions await the June 25 close). The pre-event correctly nailed the high-confidence fundamental beat but under-weighted the magnitude of the Q4 guide step-up and the SCA structural disclosure. This print also resolves the linked mu-cycle-peak-watch event BULLISH (super-cycle durable, not rolling).