Current price (2026-06-26)
$1,132
Post-earnings peak $1,255 → −9.8% from high
After the Blowout Quarter: The Earnings Call Recasts "Cycle" as "Contract," Repricing the Downside Tail
| Metric | Value | Basis / Source |
|---|---|---|
| Current price (2026-06-26 close) | $1,132.33 | tvmarket NASDAQ:MU, 6/26 −6.7% ("earnings afterglow gone," Barron's) |
| Post-earnings peak | $1,255.00 intraday / $1,213.56 close (6/25) | 6/24 earnings-day close $1,048.51 → 6/25 +15.7% |
| 52-week high / distance | $1,255.00 / −9.8% | 6/25 intraday; near all-time high |
| Market cap | ~$1.30 trillion | 1.15B diluted shares × price (6/25 peak briefly exceeded Meta and Tesla intraday) |
| Diluted shares | ~1.15B (FQ4 guidance basis) | Company guidance |
| FY26 own-quarter reconciliation | Revenue ~$129B · EPS ~$73 | Three quarters $78.96B + Q4 guidance $50B; non-GAAP EPS three quarters ~$42.09 + Q4 $31 (see §4.1) |
| Multi-method fair value midpoint | ~$1,180 | Three-scenario weighted; see §5.4 |
| Implied upside / downside (vs $1,132) | Midpoint ~$1,180 → ~+4% | Bull tail $1,850 (+63%); bear tail $520 (−54%) |
| Confidence | Medium | Bull/bear divergence >3×; SCA floors raised worst-case tail from −75% to ~−54% |
| Investment signal | Neutral | Current price approximately equals three-scenario weighted midpoint; risk/reward broadly symmetric |
Micron's FQ3 FY26 answered every remaining doubt in one report: revenue $41.5B (YoY +346%, QoQ +74%; the $17.6B sequential increase is the largest in company history), non-GAAP gross margin 84.9%, operating margin 81.2%, non-GAAP EPS $25.11, with Q4 guidance pointing to $50B / 86% gross margin / $31 EPS. Business momentum is supported entirely by cash flow — OCF $25.4B, FCF $18.3B, net cash $24.4B; all three major rating agencies upgraded to BBB+.
The structural fact that most changes the pricing framework: 16 Strategic Customer Agreements (SCAs) lock in both volume and price — the largest customer's agreement includes a ceiling price (current CQ2 market price) and a floor price running through the full contract term; management stated explicitly that the floor GM is "well above peak quarterly margins in any past cycle" (floor GM > ~60%); 14 SCAs at minimum volume × minimum price disclosed Remaining Performance Obligations of approximately $100 billion, accompanied by $22 billion in cash deposits and financial commitments (~$18 billion in cash deposits). This raises the gross-margin floor for roughly 25%–40% of revenue via contract price floors — the sharpest downside tail of a deep-cyclical stock is partially bond-ified.
Earnings remain at the absolute cycle peak (84.9% gross margin has no precedent for sustainability in memory history; management guided for "meaningful moderation in the rate of price increases" starting FQ4). The supply-response risk materializes through 2028 (FY27 capex rising QoQ, ID1 first wafer mid-CY2027, ID2 late-CY2028). Pricing summary: at $1,132, current price is approximately equal to the three-scenario weighted midpoint of ~$1,180 (bull $1,850×30% + base $1,150×45% + bear $520×25%), implying ~+4%. Risk/reward: upside tail +63%, downside tail cushioned by contract floors at −54%, broadly symmetric. Investment signal: Neutral. The structural repricing has been partially realized at current levels; the next directional signal requires industry data (SCA RPO delivery, contract price QoQ, China CXMT supply entry).
(1) FQ4 10-K discloses ~$100B RPO and next-12-month split for all 14 SCAs, verifying floor GM > historical peak; (2) DRAM quarterly contract prices hold positive QoQ (Jefferies projects CY Q3 +40–50% QoQ); (3) Apple receives US approval to source from China's CXMT (Chinese supply entry = structural negative). Any adverse crossing downgrades signal from Neutral to Bearish.
Price facts are kept separate from attribution. MU over the past 14 months has been a compounding machine where earnings followed price and price followed AI capex, but intraday volatility in recent weeks (ATR ~$95, ~8.4% of price) is at extreme levels.
| Window | Return | Anchor |
|---|---|---|
| 1 week | ≈ flat (−0.1%) | 6/18 close $1,133.99 → 6/26 $1,132.33; earnings impulse offset by pullback |
| 1 month | +26% | 5/26 close $895.88 |
| 3 months | +209% | 3/30 weekly close $366.24 |
| 6 months / YTD | +259% | 2025-12-29 close $315.42 |
| 12 months | +808% | 2025-06-23 weekly close $124.76 (~9.1×) |
Earnings-window price sequence: 6/22 close $1,211.38 (pre-earnings run-up) → 6/23 −13.18% to $1,051.77 (Korean 2× leveraged ETF forced liquidation + pre-earnings risk-off; KOSPI intraday circuit breaker; see §3) → 6/24 earnings-day close $1,048.51 → 6/24 after-hours beat → 6/25 +15.7% to $1,213.56 (intraday $1,255 all-time high) → 6/26 −6.7% to $1,132.33 (Barron's: "earnings afterglow gone"; SK Hynix/Samsung dragged KOSPI to another intraday circuit-breaker halt).
Technical levels: RSI ~59 (neutral; pulled back from 80+ during the run-up, no overbought signal); ADX ~24 and fading (trend momentum weakening); MACD positive but histogram flattening; price above the 20-day moving average at $1,035 and below Bollinger upper band at $1,217 — in a "high-altitude consolidation" zone. Summary: The earnings beat did not produce a trend breakout; instead it pushed the price to an all-time high and into wide-range consolidation — a "good-news-realized + high-level-turnover" pattern. Direction is determined by subsequent industry signals (§8).
Management disclosed the SCA structure comprehensively: 16 agreements, mostly 5-year terms (CY2026–CY2030), covering approximately 20% of DRAM volume + one-third of NAND volume; the largest customer's agreement includes a ceiling price (= current CQ2 market price) + floor price running through the full contract; upon full SCA execution, the fixed-price or ceiling-price portion represents approximately 40% of revenue; 14 SCAs at minimum volume × minimum price disclosed RPO of approximately $100 billion, plus $22 billion in cash deposits and financial commitments (~$18 billion in cash deposits).
Evidence: Sanjay's words: "the floor price corresponds to gross margins well above the peak quarterly gross margins we've had in any prior cycle"; Mark added that RPO is the "minimum enforceable amount" under ASC 606, and actual revenue "will far exceed RPO."
Implication: The prior cycle trough (FY23) saw gross margin −7.7% and loss of $4.45/share; if roughly 25%–40% of revenue has its gross margin contractually floored above historical peak cycle margins (>~60%), the normalized EPS floor for the next downturn is structurally elevated. This is the partial bond-ification of the sharpest cyclical tail, and the basis for setting the bear scenario at $520 (above the ~$260–300 for a pure-cycle trough without SCA protection).
Falsification: RPO disclosed in FQ4 10-K materially below $100B, or floor GM measured below 50%.
Management stated "no line of sight to when supply will catch up with demand," with tightness persisting beyond CY2027; CY2026 industry DRAM bit shipments +low-to-mid 20s%, NAND ~+20%, while data-center DRAM+NAND bit shipments more than doubled vs. two years ago; greenfield expansion constrained by construction lead times, skilled labor, permits, and energy infrastructure.
Evidence: FQ3 DRAM price +low-60s% QoQ, NAND +mid-80s% QoQ; industry server unit CY2026 revised up from low-double-digits to high-teens.
Implication: Supplier-led pricing in a zero-spot-market environment; management guided blended DRAM cost per bit to rise (LP5→LP6, DDR5→DDR6, new HBM generation + greenfield ramp), meaning rising costs actually support price levels and reinforce contract floor values.
Falsification: TrendForce quarterly contract price QoQ narrows to single digits or spot prices soften.
FY24 gross margin ~23% → FY25 41% → FQ3 FY26 84.9%; FQ4 guidance of 86% accompanies commentary reflecting "meaningful moderation in the rate of price increases." FY26 actual EPS ~$73, FQ4 annualized run-rate ~$124, implying forward PE ~15.5× (FY26) / ~9.1× (FY27 run-rate).
Implication: Valuation must separate "peak earnings" from "normalized earnings" (see §5); low forward PE at a cycle peak is advance pricing of earnings revisions, not a cheapness measure — this applies to the ~60% of revenue outside SCA coverage, while the SCA-floored portion is partially exempted.
Falsification: Supply discipline extends this supercycle beyond 2028, requiring upward revision to the peak assumption.
FY26 capex ~$27B (net of government subsidies); FY27 quarterly capex above FQ4 levels, with over half the YoY increase from facility construction; ID1 (Idaho) first wafer mid-CY2027, ID2 late-CY2028, New York fab broke ground in January this year.
Implication: The historical pattern of capex spikes corresponding to price peaks ~18–24 months later requires both that capex converts to effective wafer output and that demand does not simultaneously absorb new supply; HBM's rising die trade ratio continuously compresses non-HBM supply, providing a buffer.
Falsification: Actual supply/demand gap data (not the capex figure itself) turns easy.
Key Tracking Indicators (updated)Indicators: (1) ~$100B RPO delivery and floor GM landing in FQ4 10-K; (2) DRAM quarterly contract price QoQ; (3) Apple-CXMT sourcing approval.
Current levels: RPO ~$100B (FQ3-end >$5B disclosed; full 14-agreement disclosure in FQ4 10-K); DRAM contract price CY Q2 +low-60s%, Jefferies projects Q3 +40–50% QoQ; CXMT on Pentagon Entities List, no legal ban but US clearance required.
Thresholds: (1) RPO materially below $100B / floor GM <50%; (2) DRAM contract price first negative QoQ; (3) Apple clearance granted for CXMT sourcing.
Sources: MU FQ4 10-K / earnings call, TrendForce quarterly contract prices, FT/Reuters policy reporting.
Date: Next window = MU FQ4 earnings (~late September 2026) + TrendForce CY Q3 contract prices (~early October).
Implication of crossing: Any adverse threshold crossed = structural repricing falsified; Neutral downgrades to Bearish; valuation anchor reverts from "contract floor" to pure-cycle normalization.
Product lines. Micron is one of three global DRAM vertical integrated manufacturers and one of the few NAND VIMs (with SK Hynix and Samsung forming the DRAM oligopoly; NAND also includes SanDisk/Kioxia). Two technology pillars:
| Business Unit (FQ3 FY26) | Revenue | Mix | GM | QoQ | Commentary |
|---|---|---|---|---|---|
| Cloud Memory (CMBU, incl. HBM) | $13.8B | 33% | 83% | +78% | HBM primary battlefield; largest by revenue |
| Core Data Center (CDBU) | $11.5B | 28% | 87% | +103% | Enterprise DRAM/SSD; highest margin tier |
| Mobile and Client (MCBU) | $11.5B | 28% | 87% | +49% | LPDDR/UFS; volume down, price up |
| Automotive and Embedded (AEBU) | $4.6B | 11% | 79% | +71% | Most stable; least cyclical |
| Total | $41.5B | 100% | 84.9% | +74% | Four units reconcile |
Customer concentration. End-market is highly concentrated among a small number of hyperscalers and GPU platforms (HBM4 is in high-volume ramp for the NVIDIA Vera Rubin platform, ramping ~2× the speed of HBM3E 12-high). SCAs convert this concentration into a structural asset: 4 hyperscalers + 3 medium customers + smaller automotive customers, bound by take-or-pay volume commitments, with $18 billion in cash deposits — customers paying real cash to reserve 5-year supply, the strongest evidence of pricing power and demand visibility.
Revenue and margin logic. Memory is commodity-priced, but HBM/premium DRAM has periodic pricing power due to technical barriers + capacity binding. Margin logic = (contract price − cash cost) × bit shipments; in shortage + price-increase cycles, price elasticity far exceeds cost, driving gross margin up non-linearly (the source of the 23%→85% move). The SCA price band holds the lower half of this curve (cycle downturn) up with a floor, and gives up the upper half (cycle upturn) above the ceiling to customers — trading "upside concession" for "downside certainty."
Capital intensity and operating leverage. Extremely capital-heavy: FY26 capex ~$27B (net of subsidies); FY27 rising per quarter. The two-sided nature of operating leverage was fully demonstrated this quarter (revenue +74% QoQ while opex rose only $97M → operating margin 81.2%); in downturns, fixed D&A rapidly compresses gross margin — which is exactly the risk the SCA floor is designed to hedge.
Structural changes (disclosed facts): (1) Segment framework switched to four end-markets; (2) SCA coverage target "≥50% of revenue," currently ~20% DRAM + 1/3 NAND volume; (3) US domestic capacity (Idaho ID1/ID2, New York) + CHIPS Act (from 2026-12-09 plans to raise capital return; long-term "100% of excess cash" to shareholders); (4) Multi-year EUV supply agreement with ASML (1-delta node), locking in advanced process capacity.
Value-chain position. Micron sits in the "memory device manufacturing" layer; upstream are WFE equipment (ASML/AMAT/Lam) and silicon wafers; downstream are GPU/ASIC makers (NVIDIA/AMD/Broadcom/Google TPU) and OEMs. HBM has partially elevated Micron from "commodity cycle supplier" to "critical AI compute component supplier," though the fundamental nature remains oligopoly manufacturing.
| Company | Role | Current position | Comparability note |
|---|---|---|---|
| SK Hynix (000660.KS) | HBM leader | HBM share ~58–62%; 12M significantly outperforming | Most directly comparable; HBM leader; MU's share benchmark |
| Samsung (005930.KS) | DRAM/NAND full-line | HBM ~17–21%; conglomerate SOTP discount | Comparable but diversified; valuation dragged by non-semiconductor; $648B 10-year investment plan = supply-response signal |
| SanDisk (SNDK) | Pure NAND | Last quarter revenue +251% YoY; GM ~78% | Pure-play NAND; validates NAND cycle; no DRAM/HBM exposure |
| CXMT (China, unlisted) | Niche DRAM challenger | Apple seeking US approval to source (FT 6/27) | Structural supply variable; see "misread" paragraph below |
Why peer multiples mislead: SK Hynix, Samsung, and SanDisk all show "low forward PE" because the entire industry is at peak earnings — the denominator (E) is at an unsustainable high. A cross-sectional PE comparison with all three at peak E means all three look "cheap" — this reflects cyclical synchrony, not independent safety margin for MU. MU is a share follower in HBM (~21% vs. SK Hynix ~60%), but the earnings call clarified this is a deliberate choice: "We intentionally keep our HBM share in line with our DRAM share because HBM consumes large wafer capacity and compresses non-HBM supply" — using share capping as a capacity-allocation tool.
Market misread identification: Markets attributed the 6/26 sell-off primarily to "Micron itself peaking," but the dominant driver was Korean peer (SK Hynix/Samsung) sentiment contagion dragging KOSPI to another intraday circuit-breaker halt + high-level stock turnover, not MU fundamental deterioration. A second misread treats Apple-CXMT as a near-term negative: in the short term CXMT can only supply niche/low-end DRAM with no HBM/premium DRAM replacement capability; but as a structural medium-to-long-term variable (Chinese supply entry + Pentagon Entities List policy uncertainty), it is a real signal that must be included in the bear scenario.
| Channel | Category | Mechanism | True weight for MU pricing |
|---|---|---|---|
| HBM/DRAM contract prices, supply/demand gap | Direct (fundamental) | Same oligopoly pricing system; contract prices move together | High — same industry fundamentals |
| Korean 2× leveraged ETF forced liquidation, KOSPI circuit breaker | Sentiment (contagion) | Leveraged position unwinding spills to global high-beta semis | Medium — amplifies near-term volatility; does not change fundamentals |
| Samsung $648B expansion, CXMT entry | Indirect (supply response) | Supply-response expectations; 18–24 month lag | Medium-to-long-term high — primary driver of bear scenario |
The FY26 stale consensus still circulating ($108.7B revenue / $58 EPS) is severely outdated. Reconciled using company quarterly actuals:
| Quarter | Revenue | Non-GAAP EPS | Notes |
|---|---|---|---|
| FQ1 FY26 | $13.64B | ~$4.79 | Realized |
| FQ2 FY26 | $23.86B | ~$12.19 | Realized |
| FQ3 FY26 | $41.46B | $25.11 | Realized; +106% QoQ |
| FQ4 FY26 (guidance) | $50B ±$1B | $31 ±$1 | Company guidance |
| FY26 total | ~$129B | ~$73 | ~+19% / +26% above stale consensus |
FQ4 annualized run-rate = $31 × 4 = $124. Any forward multiple calculated with "$58 EPS × PE" is wrong; the true peak EPS anchor is ~$73 (FY26) or annualized $124.
| Line item | FQ3 FY26 | QoQ | YoY | Commentary |
|---|---|---|---|---|
| Revenue | $41.5B | +74% | +346% | 5th consecutive record; $17.6B sequential gain is all-time largest |
| DRAM revenue | $31.3B | +67% | +343% | 76% of total; price +low-60s%, volume +low-single-digits |
| NAND revenue | $9.9B | +99% | +361% | 24% of total; price +mid-80s%, volume +mid-single-digits |
| Gross margin (non-GAAP) | 84.9% | +10pp | ×2+ | Company record |
| Opex | $1.5B | +$97M | — | Rose only with variable compensation; full operating leverage realized |
| Operating income | $33.7B | — | — | Operating margin 81.2% (+12pp QoQ, +54pp YoY) |
| Tax | $5.1B | — | — | Effective tax rate 14.9% |
| Non-GAAP EPS | $25.11 | +106% | — | Above the high end of guidance |
Earnings-to-cash quality (strong): OCF $25.4B, capex $7.1B, FCF $18.3B (quarterly record); profit and cash flow move in tandem. Inventory $8.6B, DIO 120 days; DRAM inventory <120 days and "very tight" — inventory tightness is consistent with the pricing narrative.
Cash and investments record $30.2B; retired $4.4B of debt this quarter (including $4.3B tender offer on senior notes); period-end total debt only $5.7B, net cash $24.4B; all three major rating agencies upgraded to BBB+. SCA cash deposits (~$18B; ~$10B more expected in FQ4) flow through financing activities and do not affect FCF; they are unrestricted cash, returned to customers in the contract back half. Capital return: from 2026-12-09 (2nd anniversary of CHIPS Act) the company plans to raise returns, with management stating a long-term commitment to return "100% of excess cash to shareholders."
With earnings at a physical cycle peak, single-multiple approaches are misled by peak E. This report anchors on scenario-weighted valuation (explicitly modeling "how long can the cycle last"), uses peer multiples as cross-sectional validation, and treats DCF as sensitivity-range-only. The introduction of SCA floor prices requires using "contract-floored normalized EPS" rather than bare-cycle trough EPS in the bear scenario.
| Company | Forward PE (peak E) | Notes |
|---|---|---|
| MU | ~15.5× (FY26 $73) / ~9.1× (FY27 run-rate $124) | Peak E; low PE is a cycle-peak signal; SCA floor partially exempt |
| SK Hynix | Peak-basis ~10× / normalized ~30–40× | Divergence >3×; same valuation trap structure |
| SanDisk | NTM ~11–12× | Pure NAND; peak E |
Cross-sectional conclusion: all three looking "cheap" reflects cyclical synchrony; MU's SCA price band is the most systematic among peers (16 agreements, ~$100B RPO), giving it relatively thicker "floor protection."
| Scenario | Prob. | Logic | Anchor EPS | Multiple | Target |
|---|---|---|---|---|---|
| Bull | 30% | SCA + structural shortage persists; CY2027 prices rise further (Jefferies Q3 +40–50%) and hold; cycle "contracted" into growth stock | ~$150–180 (CY2027) | ~12–14× | $1,850 |
| Base | 45% | FY26 EPS ~$73 / FY27 ~$130 run-rate; cycle-perceived multiple; SCA floor supports slightly higher trough multiple vs. pure cycle | ~$120–130 (FY27) | ~9–10× | $1,150 |
| Bear | 25% | 2027–28 supply response + non-SCA revenue price decline; SCA floor (GM >60% covering ~25–40% of revenue) holds normalized EPS | ~$30–40 (normalized) | ~13–15× | $520 |
Bull target $1,850 aligns with the most aggressive sell-side targets (Melius $2,200, HSBC $1,700); bear target $520 is materially above the pure-cycle normalized floor of ~$260–300; the difference is the value of the SCA contract floor.
Weighted = 0.30 × $1,850 + 0.45 × $1,150 + 0.25 × $520 = ~$1,180. Implied upside vs. current price $1,132: ~+4%.
Deep-cyclical DCF is highly sensitive to the normalized FCF and fade path; used only to bracket boundaries. Base assumptions: FY27 FCF run-rate ~$40–60B (peak; not extrapolable), fading to a normalized mid-cycle level of ~$12–18B/year; WACC 10.5%, terminal growth 3%, net cash $24.4B, diluted shares 1.15B. On a "peak FCF for 2 years then fade to normalized" path, DCF fair-value range is approximately $900–1,400/share (terminal value >70% of total value; highly sensitive to fade assumptions; not used as primary anchor). DCF and scenario-method midpoint ($1,180) are in the same order of magnitude, providing cross-validation.
Bull ($1,850) vs. bear ($520) divergence >3×; structural divergence between peak-basis and normalized-basis methods persists — per methodology this warrants reducing confidence by one notch. SCA floors raised the worst-case tail from "normalized ~$300, −75% from spot" to "$520, −54%," narrowing downside uncertainty. Summary: Confidence: Medium; current price ≈ weighted midpoint; risk/reward broadly symmetric; signal Neutral.
Form 4 activity in the past 60 days (May–June 2026): multiple transactions are routine RSU vesting/tax withholding and 10b5-1 plan sales (filed 5/01, 5/11, 5/29, 6/09), accompanied by two Form 144 prospective-sale filings (5/11, 5/29). Against the backdrop of the stock rising from ~$540 (late April) to ~$1,200 (June), executive sales under 10b5-1 plans are pre-programmed and must be distinguished from discretionary open-market selling. Current public data shows no large discretionary open-market sales by senior executives (CEO/CFO), though specific "transaction code" classification for each Form 4 requires line-by-line verification in the FQ4 10-K window (noted as "not individually verified"). Governance positives: all three rating agencies upgraded to BBB+; clear "100% excess cash return" path stated.
| Metric | FQ4 FY26 Guidance | Commentary |
|---|---|---|
| Revenue | $50B ±$1B (record) | Sequential +20% |
| Gross margin | ~86% | Record; management flagged "meaningful moderation in rate of price increases" |
| Opex | ~$1.65B | FY27 full-year opex +$1B (R&D; skewed to 2H) |
| EPS | $31 ±$1 (record) | ~1.15B diluted shares |
| Tax rate | ~15% | — |
| Capex | ~$10B (FQ4) → FY26 ~$27B | FY27 quarterly capex above FQ4; over half the YoY increase from facility construction |
Management qualitative commentary (earnings call): shortage persists beyond CY2027; no line of sight to supply catching demand; blended DRAM cost per bit to rise; HBM4 >$1B shipped and 12-high ramp 2× faster than HBM3E; next-generation node in high-volume production 2H CY2027; Tongluo (Taiwan) production line shipping mid-CY2027 (one quarter ahead); Singapore becoming HBM advanced packaging hub (1H CY2027).
Results that would confirm the thesis: FQ4 full RPO disclosure ~$100B + floor GM landing; DRAM contract price Q3 continues +40–50%.
Results that would falsify the thesis: Price deceleration faster than expected; RPO below stated range; CXMT approval granted.
| Catalyst | Probability | Impact | Watch signal |
|---|---|---|---|
| FQ4 RPO ~$100B delivery + floor GM confirmed | High | High | FQ4 10-K next-12M split |
| CY Q3 DRAM contract price +40–50% (Jefferies) | Medium-high | High | TrendForce ~early October |
| HBM4 share/revenue upside surprise | Medium | Medium | FQ4 HBM revenue disclosure |
| Capital return acceleration (from 12/09) | High | Medium | Buyback/dividend announcement |
| Risk | Probability | Impact | Watch signal |
|---|---|---|---|
| 2028 supply response peaks contract price (ID1/ID2 ramp + Samsung $648B) | Medium | High | Supply/demand gap data turning easy; contract price first negative QoQ |
| Apple-CXMT approved; Chinese supply enters market | Medium-low | Medium-high | FT/Reuters policy; US ruling |
| Non-SCA revenue (~60%) sold down with cycle | Medium | High | Spot/contract prices; inventory DIO rising |
| Peak EPS repriced by market as cycle top | Medium | High | Forward PE compression; valuation derate |
| Single hyperscaler GPU roadmap delayed | Low | Medium-high | NVIDIA/hyperscaler capex guidance |
SCA + shortage extends to 2028+, SCA mix reaches 50%, EPS holds at high levels; cycle structurally extended.
Price increase rate moderates post-FQ4 but remains elevated; FY27 EPS ~$130; SCA floor makes downturn gradual.
2H 2027 supply response + non-SCA price decline; SCA floor holds normalized EPS ~$30–40.
| Assumption | Value | Notes |
|---|---|---|
| FY27 FCF (peak) | $40–60B | Not extrapolable; upper-bound only |
| Normalized mid-cycle FCF | $12–18B/year | SCA floor raises the lower bound |
| Forecast period | 10 years | Fade from peak to normalized |
| WACC | 10.5% | Deep cyclical; high beta |
| Terminal growth | 3% | Long-term AI demand |
| Net cash | $24.4B | FQ3 end |
| Share count | 1.15B | FQ4 guidance |
| DCF fair-value range | $900–1,400 | Terminal value >70%; highly fade-sensitive; boundary use only |
Sensitivity: The fade path is the dominant variable — one additional year of peak FCF pushes the DCF upper bound toward $1,500+; reversion starting 2027 pushes the lower bound toward $800. This divergence has the same root cause as the scenario-method bull/bear gap. DCF and scenario-method midpoint ($1,180) are in the same order of magnitude, providing cross-validation.
Produced by NewsLiquid; all data and conclusions are independent research outputs. Deep-cyclical valuation depends heavily on the unknowable variable of "cycle duration," and peak earnings are not linearly extrapolable. This report is not investment advice.