Micron & the Supercycle — Faircurve Equity Pulse · 1 Jun 2026

Micron closed at $971, up nearly tenfold in twelve months and roughly +75% in May alone, on the sharpest memory up-cycle on record. The earnings are not in dispute: fiscal Q2 delivered $23.9B of revenue, a 75% gross margin and $12.08 of EPS, with DRAM prices up mid-60s% sequentially as HBM for AI accelerators met deliberately tight supply. The call turns on one distinction. AI demand looks structural — hyperscaler capex is committed years out in the trillions, high-bandwidth memory (HBM) is sold forward under multiyear agreements, and its supply is gated by advanced packaging rather than easily-added wafers — and Faircurve’s base credits that, valuing the HBM-era run-rate as the floor, far above any prior cycle. What it does not extrapolate is the pricing layer: a 75% gross margin guided to 81%, on selling prices the Street’s own consensus expects to fade. Stripping the pricing spike while crediting the structural demand places base fair value at $640; if pricing proves as durable as demand, the bull case is $1,150, modestly above spot. The initiation is SELL · $640 (−34%), at reduced conviction — a call on whether the pricing spike, not the demand, is permanent.

By Faircurve Research

Micron & the Supercycle

Faircurve Equity Pulse · Single-Name Research
NASDAQ : MU1 Jun 2026
SPOT $971.00 (+5.1%)
SELL · $640

Micron closed at $971, up nearly tenfold in twelve months and roughly +75% in May alone, on the sharpest memory up-cycle on record. The earnings are not in dispute: fiscal Q2 delivered $23.9B of revenue, a 75% gross margin and $12.08 of EPS, with DRAM prices up mid-60s% sequentially as HBM for AI accelerators met deliberately tight supply. The call turns on one distinction. AI demand looks structural — hyperscaler capex is committed years out in the trillions, high-bandwidth memory (HBM) is sold forward under multiyear agreements, and its supply is gated by advanced packaging rather than easily-added wafers — and Faircurve’s base credits that, valuing the HBM-era run-rate as the floor, far above any prior cycle. What it does not extrapolate is the pricing layer: a 75% gross margin guided to 81%, on selling prices the Street’s own consensus expects to fade. Stripping the pricing spike while crediting the structural demand places base fair value at $640; if pricing proves as durable as demand, the bull case is $1,150, modestly above spot. The initiation is SELL · $640 (−34%), at reduced conviction — a call on whether the pricing spike, not the demand, is permanent.

§ 01 — Tape & Setup

The Snapshot

PRICE
$971.00
+5.1% 1d
MARKET CAP
$1.10T
+889% 1Y
12M TARGET
$640
-34% downside
RATING
SELL
Faircurve
FWD P/E (FY27)
9.6x
peak earnings · 45x trailing
GROSS MARGIN
74%
Q2; guided ~81%
FCF YIELD
0.9%
trailing; capex >$25B
REV GROWTH FY27E
+64%
cons; FY26 +190%
MU Micron Technology, Inc. 1-YEAR PRICE HISTORY $971.00 $200 $400 $600 $800 $1000 52W HIGH $971 2026-05-29 · spot 52W LOW $98.18 2025-06-02 Jul 25 Oct 25 Jan 26 Apr 26 SPOT PRICE 50DMA 200DMA 1Y RETURN +889% MAX DRAWDOWN -30% 30D VOL (ANN.) 88% BETA (5Y) 1.92
§ 02 — Thesis · Bull vs Bear

The Debate at $971

HBM is a genuine step-change

High-bandwidth memory is sold out through calendar 2026 under multiyear agreements and prices more like a custom-logic product than a spot commodity — structurally lifting the mid-cycle margin floor.

A disciplined three-player oligopoly

DRAM is now Micron, SK Hynix and Samsung. Capital discipline and surging AI demand have kept the market in deficit for six consecutive quarters.

AI demand is secular, not a quarter

Data-center is more than half of revenue; accelerator and high-capacity memory module bit-demand compounds with every training and inference cycle.

Record cash, net-cash balance sheet

Operating cash flow of $11.9B in the quarter funds a >$25B capex build and a raised dividend without leverage.

Peak margins always mean-revert

A 75% gross margin — guided to ~81% — is unprecedented for memory and has no durable historical analogue. The question is when pricing normalises, not whether.

Supply is being built into the peak

Micron is lifting FY26 capex above $25B while Hynix and Samsung expand. Memory down-cycles are made by the capacity ordered at the top.

A low multiple on peak earnings is the trap

Memory has historically been cheapest on P/E at the top. 9.6x forward is not a discount but a signal the market is not extrapolating the peak.

Elasticity is already biting demand

Management flags that memory pricing may crowd out content, with PC and smartphone units at risk of low-double-digit declines in calendar 2026.

Faircurve’s read

The franchise debate is settled; the valuation debate turns on one distinction. AI demand durability looks genuinely structural — hyperscaler capex is committed years out in the trillions, high-bandwidth memory (HBM) is sold forward under multiyear agreements, and its supply is gated by advanced packaging rather than easily-added wafers. Faircurve credits that: the base values the FY26 HBM-era run-rate as the floor, far above any prior cycle. What it does not extrapolate is the pricing layer on top — a 75% gross margin guided to 81%, with DRAM selling prices +mid-60s% in a single quarter. Strip the spike, keep the structural demand, and fair value is ~$640. The call is whether the spike, not the demand, is permanent.

Three things to watch

Three signals settle the durability question: the DRAM pricing trajectory into the seasonally softer first half of calendar 2026; industry capex additions from all three makers, the mechanism that has ended every prior up-cycle; and the breadth of HBM demand beyond the current accelerator roadmap. A second consecutive quarter of sequential price gains would push the read toward the bull case; the first sign of inventory rebuild at hyperscalers would confirm the base.

§ 03 — The Latest Print

Q2 FY26 — the peak, in numbers

Micron’s February quarter was not merely a record; it was a different order of magnitude. Revenue of $23.9B rose +196% year-on-year and +75% sequentially, as DRAM contract prices climbed mid-60s% in a single quarter and HBM shipped at volume into the latest accelerator roadmap. The gross margin reached 74.4% — a level memory has never sustained — and management guided the next quarter higher still, to roughly 81%. The signal worth holding is precisely that: a margin this far above the franchise’s own history is the definition of a cyclical peak, not a new normal.

Revenue $23.9B (+196% year-on-year, +75% sequential) — a record by a wide margin; DRAM $18.8B (79% of revenue), NAND $5.0B, with DRAM prices up mid-60s% sequentially.
Gross margin 74.4%; guided ~81% next quarter — an unprecedented level for memory and the clearest evidence the cycle is at a peak, not mid-cycle.
Diluted EPS $12.08; operating cash flow $11.9B; free cash flow $6.9B — record cash generation even against a >$25B capex build.
FY26 capex raised above $25B; first five-year supply agreement signed; dividend raised 30% — management is investing into the peak and locking multiyear HBM commitments.

Operational dashboard, Q2 FY26 (quarter ended 26 Feb 2026)

MetricQ2 FY26SequentialYear-on-year
Revenue$23.86B+74.9%+196.3%
DRAM revenue$18.8B (79%)strongrecord
NAND revenue$5.0B (21%)strongrecord
Gross profit$17.75B
Gross margin74.4%+18 pts+37 pts
Operating margin67.6%
Diluted EPS$12.08+163%+757%
Operating cash flow$11.9B+42%record
Free cash flow$6.9B+77%record
Capex (quarter)$5.0B

Eight-quarter trajectory — revenue & non-GAAP gross margin

Q3'24Q4'24Q1'25Q2'25Q3'25Q4'25Q1'26Q2'26
Revenue ($B)6.87.88.78.19.311.313.623.9
Gross mgn (NG)28%36%38%38%39%46%57%75%
EPS (NG, $)0.621.181.791.561.913.035.2012.08
§ 04 — Five-Year Trends

The cyclicality, in one table

Fiscal year (Aug)RevenueGross mgnEBITDANet incomeDil EPSCapexFCF
FY21$27.7B37.6%$12.9B$5.9B$5.14$10.0B$2.4B
FY22$30.8B45.2%$16.7B$8.7B$7.74$12.1B$3.1B
FY23$15.5B-9.1%$2.2B-$5.8B-$5.34$7.7B-$6.1B
FY24$25.1B22.4%$8.9B$0.8B$0.70$8.4B$0.1B
FY25$37.4B39.8%$18.5B$8.5B$7.59$15.9B$1.7B
FY26E (cons)$108.6B~70%$58.87>$25B

No table makes the case more economically than this one. In the space of three fiscal years Micron swung from a $8.7B profit (FY22) to a $5.8B loss with a negative gross margin (FY23) and back to record profitability — the signature of a commodity whose price, not its volume, drives the income statement. The current cycle is larger in amplitude than any before it: FY25’s $7.59 of EPS is on track to multiply roughly eightfold to a consensus $58.87 in FY26 and above $100 in FY27. That step-change is real, and HBM is a genuine reason the trough should be higher than in the past. It is not a reason to treat the peak as the baseline — the FY23 column is only thirty months old.

§ 05 — Earnings Call Signal

What management said, eight quarters running

Theme frequency — mentions per call (oldest → newest)

ThemeQ3'24Q4'24Q1'25Q2'25Q3'25Q4'25Q1'26Q2'26
HBM2724173238333024
Data center / AI4441304749454246
DRAM3834244136393331
Supply / tightness1618131719243127
Pricing / selling prices1416111917182119
Capex / capacity2224121918212725
Inventory111318128644
Long-term / supply agreements31211249

Recurring metrics management quotes

MetricQ3'24Q2'25Q4'25Q1'26Q2'26
Total revenue$6.8B$8.1B$11.3B$13.6B$23.9B
Non-GAAP gross margin28%38%46%57%75%
DRAM price (sequential)+~20%+mid-s.d.+low-d.d.+20%+mid-60s%
NAND price (sequential)+~20%-high-teens+high-s.d.+mid-teens+high-70s%
HBM run-rate>$100M/q>$1B/q~$8B ann.recordrecord
Free cash flow$0.4B$0.9B$0.8B$3.9B$6.9B
What this tells us

The transcript arc is a near-monotonic march from recovery to euphoria: gross margin from 28% to 75%, HBM from a sub-$100M curiosity to an ~$8B annualised run-rate, and DRAM pricing re-accelerating to +mid-60s% in a single quarter. Two sub-surface signals matter for the call. First, the “inventory” line fades to near-silence while “supply / tightness” and “supply agreements” rise — the language of a peak. Second, management itself frames pricing as structurally driven yet concedes elasticity risk in PCs and phones. Both are consistent with a cycle that is extraordinarily strong and, by management’s own framing, extraordinary rather than normal.

§ 06 — Earnings Call Signal · Voices

Guidance cadence and three quotes

Next-quarter guidance track

Guided forRev midpointGM guideEPS guideFraming
after Q3'24$7.6B34.5%$1.08HBM sold out calendar 2024/25
after Q2'25$8.8B36.5%$1.57HBM >$1B; tight leading edge
after Q4'25$12.5B51.5%$3.75records; HBM share = DRAM share
after Q1'26$18.7B68%$8.42calendar 2026 sold out; tight beyond calendar 2026
after Q2'26$33.5B~81%$19.15first 5-yr supply agreement; supply far below demand
Sanjay Mehrotra · Chairman, President & CEO · Q3 FY24 · FUTURE-VISION
“We are in the early innings of a multi-year race to enable artificial general intelligence… These trends will drive significant growth in the demand for DRAM and NAND, and we believe that Micron will be one of the biggest beneficiaries in the semiconductor industry of the multi-year growth opportunity driven by AI.”
Mark Murphy · CFO · Q1 FY25 · CANDID-DOWNSIDE
“The NAND challenging market conditions, we believe, extend into calendar Q1… and then… the underload charges will begin to hit… in the form of period costs and then just some higher-cost inventory.”
Sanjay Mehrotra · Chairman, President & CEO · Q1 FY26 · CONDITIONAL-BULL
“Sustained and strong industry demand, along with supply constraints, are contributing to tight market conditions, and we expect these conditions to persist beyond calendar 2026. We are making progress with customers in our multiyear contracts with specific commitments.”
Reading the arc

Guidance has gone vertical — from a $1.08 EPS guide two years ago to $19.15 for the coming quarter — and the qualitative framing has hardened from “sold out” to “supply much less than demand.” That is the bull case in management’s own words. The one consistent counterweight, voiced by the CFO, is that memory remains a commodity: NAND weakness, underload charges and elasticity have all surfaced even inside this up-cycle. The durability question is whether the multiyear agreements convert “sold out” into a genuinely higher floor — or merely delay the point at which added supply meets cooler demand.

§ 07 — Peer Set · Forward Multiples

Where it trades vs the complex

Faircurve-curated peer set. Micron is anchored against the memory & storage complex that shares its narrative engine and cycle: the DRAM oligopolists SK Hynix and Samsung Electronics, the NAND pure-play SanDisk, and HDD-maker Western Digital as a storage-cycle adjacency. All multiples use each name’s next-fiscal-year (FY+1) consensus, every name on a price-over-consensus basis. Anchor years: MU FY27 (Aug’27), Hynix & Samsung FY27 (Dec’27), SanDisk & WDC FY27 (Jun’27) — all ≈ calendar 2027. Korean caps converted at ₩1,410/$. The crucial read: the DRAM/NAND makers all trade at single-digit forward P/Es — SK Hynix 6.2x, Samsung 6.5x, SanDisk 9.6x, Micron 9.6x — while HDD (WDC, 30.6x) sits much higher because it is earlier in its own ramp. Because two of four peers screen n/m on forward EV/EBITDA (SanDisk’s GAAP EBITDA is negative on post-spin amortisation), the valuation blend substitutes forward EV/Sales for the EV/EBITDA leg.

Forward P/E (FY+1) — sorted high to low

WDC
30.6x
Peer avg
13.2x
MU
9.6x
SanDisk
9.6x
Samsung
6.5x
SK Hynix
6.2x

Full peer table — forward FY+1 multiples

TickerMkt capFwd P/EFwd EV/EBITDAFwd P/SFwd EBITDA mgnFwd rev gr
SK Hynix≈$1.20T6.2x8.2x3.8x45.8%+37%
Samsung Elec≈$1.64T6.5x10.5x2.9x26.5%+21%
SanDisk$251B9.6xn/m5.9xn/m+117%
Western Digital$183B30.6xn/m10.4xn/m+37%
Peer avg (ex-MU)13.2x9.4x5.7x36.2%+53%
MU$1.10T9.6x12.3x6.1x50.0%+64%

Memory-maker cluster only (Hynix, Samsung, SanDisk), the truest comparison for cycle phase: forward P/E mean 7.4x. Micron at 9.6x sits at the rich end of the DRAM/NAND oligopoly — there is no relative-value discount to lean on. SanDisk’s +117% forward revenue growth and negative GAAP EBITDA reflect a noisy post-spin estimate base and are footnoted as such; WDC’s 30.6x is an HDD name earlier in its own up-cycle, not a memory comp.

Competitive-position read

On its merits, Micron is a co-leader of a consolidated, structurally improved industry, with the strongest balance sheet of the three DRAM makers and a credible HBM franchise. The peer set does not, however, supply a valuation cushion. Every memory maker trades on a single-digit forward P/E, and Micron at 9.6x is the most expensive of them. On Micron specifically the low multiple is unambiguously a peak multiple: the Street’s own consensus has EPS cresting in FY27 (~$101.58) and fading to ~$69 by FY29 — so 9.6x is paid on earnings analysts already expect to decline. The relative-value case for the stock would require the peak to hold — the proposition the absolute valuation tests.

§ 08 — Valuation

The math, line by line

Base-case derivation (normalised through the cycle)

VariableFormula · inputs · arithmeticBase value
Normalised EPSFY26 HBM-era run-rate (cons $58.87) held as the structural floor — AI-demand step-up credited, FY27 pricing spike ($101.58) stripped$58
Mid-cycle P/Ememory makers 6–10x on peak EPS → ~12.5x on normalised EPS (prior mid-cycles 10–14x; HBM-leader premium)12.5x
→ P/E fair value$58 × 12.5x$725
Normalised revenue≈ FY27 cons $178.5B less pricing-spike premium~$175B
Fwd EV/Sales (sub.)memory-maker fwd mean ~4.2x, applied just below peak4.0x
→ EV/Sales fair value$175B × 4.0x = $700B EV + $2B net cash ÷ 1.14B sh$616
DCF — discount rate (WACC)Cost of equity = 4.3% risk-free + (1.92 beta × 5.0% equity risk premium) = 13.9%; blended with 4.7% after-tax cost of debt (debt ≈1% of capital)13.8%
DCF — terminal growthnominal-GDP-anchored3.0%
→ DCF fair value5-yr FCF fade off the super-cycle at a 13.8% discount rate~$540
Blended fair value40% P/E ($725) + 40% EV/Sales ($616) + 20% DCF ($540)$644 → $640

Bull / bear flex

VariableBEARBear: whyBASEBULLBull: why
Normalised EPS$32selling prices crack as supply lands; margins revert, though the HBM floor stays above prior cycles$58$98pricing proves as durable as demand; earnings plateau near FY27–28 consensus
Mid-cycle P/E12.5xtrough on low EPS12.5x11.7xdurable plateau warrants a near-peak multiple
Fair value$400$640$1,150
§ 09 — Valuation · Methodology + Scale

Blended fair value & what’s priced in

$971 SPOT $400 BEAR -59% $640 BASE -34% 12M TARGET $1,150 BULL +18%

The valuation deliberately breaks from the standard forward template because Micron is a commodity cyclical at a margin peak, where applying any multiple to peak earnings endorses the top. The fix is to value normalised earnings power — the HBM-era FY26 run-rate (~$58 EPS) held as the floor, which already credits a structurally higher AI-demand base (the prior cycle earned $7.59) and strips only the FY27 pricing spike ($101.58) — at a mid-cycle multiple, substituting forward EV/Sales for EV/EBITDA where peers screen n/m. The 40/40/20 blend yields $644, rounded to a $640 target. Independent anchors corroborate the direction: FMP’s reference levered DCF prints $386; sell-side consensus PT is $625 (median $450; range $310–$1,625) — both below spot, though both may lag a stock that roughly doubled in May. Even probability-weighting the three cases (25% bear / 45% base / 30% bull) leaves fair value near $730 (−25%).

What the market is pricing in

At today’s $971ImpliedVersus
Revenue growth priced in (5y to a 10% return, exit 4.0× EV/Sales)~+25%/yrFY27→FY31 ($178B→~$440B, 2.5×); needs pricing to hold — bit demand alone is only ~15–20%/yr
Implied sustainable EPS at 12.5×~$78~76% of the $101.58 FY27 peak, held permanently
EPS growth priced in (same basis, exit 12.5×)~+5%/yrflattered by peak margins — ignore in favour of the revenue line
P/E paid on FY27 peak EPS ($101.58)9.6xvs 5–8x historic memory peak multiple
The proposition, reverse-engineered

Reverse-engineering the price is the cleanest test of what is embedded. The single most judgeable figure is on the revenue line, which strips the peak-margin illusion: to clear a 10% annual return over five years — exiting at a normalised 4.0× EV/Sales — Micron’s revenue must compound at about +25%/yr, from ~$178B (FY27) to ~$440B by FY31. Against long-run DRAM bit demand of ~15–20%, that requires pricing to hold near the spike, not merely volumes to grow. In earnings terms the same price implies a sustainable EPS of ~$78 — roughly three-quarters of the FY27 peak held in perpetuity, and well above the ~$58 structural-demand floor the base credits; paying 9.6x the FY27 peak, when memory has historically commanded 5–8x at cyclical tops, means the market is treating the pricing spike, not merely the demand, as semi-permanent. The stock is not pricing the AI build-out so much as the build-out at today’s prices, without a down-cycle. That is the proposition Faircurve declines near $971 and would accept much closer to the $640 base — and the one the bull case, if pricing proves as durable as demand, renders roughly fair at $1,150.

§ 10 — The Range · 12-Month View

Three cases, one call

BEAR · 25%
$400
-59% vs spot
Pricing reverts as supply lands: industry capacity additions meet cooler demand, DRAM/NAND selling prices crack, margins normalise. Normalised EPS settles near $32 — still far above prior-cycle troughs thanks to the HBM floor — at ~12.5x. In line with FMP’s DCF and the low end of the sell-side range.
BASE · 45%
$640
-34% vs spot
The AI-demand step-up holds (FY26 HBM-era run-rate, ~$58 EPS, as the structural floor) but the FY27 pricing spike normalises. ~$58 at ~12.5x, blended with EV/Sales and a high-discount-rate DCF, gives $644. A business permanently better than past cycles, valued on its durable earnings rather than the peak.
BULL · 30%
$1,150
+18% vs spot
The structural case proves right on pricing as well as demand: hyperscaler capex stays committed, HBM supply stays gated by packaging, the three makers hold discipline, and margins plateau near the peak — earnings hold ~$98 (≈ FY27–28 consensus) at ~11.7x. The ‘this time is different’ outcome; at +18% it is the price of being right, not a windfall.
SELL — 12-month price target $640
Faircurve initiates Micron at SELL with a $640 twelve-month target (about −34%), at reduced conviction. This is not a franchise call, nor a call against the AI build-out: hyperscaler capex commitments are vast and multiyear, and the base already treats the HBM-era demand step-up as structural, valuing the FY26 run-rate (~$58 EPS, far above prior cycles) rather than haircutting it. The SELL rests on a narrower proposition — that the pricing layer (a 75% gross margin guided to 81%, selling prices +mid-60s% in a quarter) normalises, as memory pricing always has, even when demand stays strong. Reverse-engineered, $971 prices ~76% of the FY27 peak EPS as permanent and pays 9.6x earnings the Street’s own consensus expects to fade to ~$69 by FY29. Conviction is deliberately moderate and the asymmetry two-sided: if HBM pricing proves as durable as the demand, the bull case is fair at $1,150 (+18%), and even probability-weighting a 30% structural outcome leaves fair value near $730 (−25%). Faircurve would move to HOLD on a second consecutive quarter of held or rising DRAM pricing — the cleanest evidence the spike is a plateau — and turn constructive in the mid-$700s.
§ 11 — Risks & Catalysts

What breaks the thesis, what triggers the move

Six principal risks (to our SELL)

Pricing durability (this time is different)
HIGH
The base credits structural AI demand but assumes pricing normalises. If HBM’s packaging-gated supply and the multiyear agreements hold margins near current levels — against trillions in committed hyperscaler capex — normalised earnings are far higher and the bull case ($1,150) becomes the base. The single biggest risk to a SELL.
AI-driven re-rating / momentum
HIGH
The entire memory complex (Hynix, Samsung, SanDisk, WDC) has re-rated together. A continued rally in AI-levered shares can sustain peak multiples and carry Micron well past fair value, irrespective of normalised earnings.
Pricing stays tight longer
MED
If DRAM/NAND contract prices hold or rise into the first half of 2026 rather than softening, FY27 earnings could exceed even the $100-plus consensus and delay any reversion.
Consensus keeps trailing the run-rate
MED
Sell-side price targets (median $450) plainly lag a stock that has roughly doubled in a month; a wave of upward target revisions could re-anchor sentiment higher near-term.
Capital return acceleration
MED
Record free cash flow and a net-cash balance sheet could fund buybacks and a faster-rising dividend, compounding per-share value through the peak.
Supply discipline beyond Micron
LOW
If Hynix and Samsung hold capex discipline rather than chasing the peak, the down-cycle is milder and later than memory’s history would imply.

Four near-term catalysts

  1. Q3 FY26 results (late June 2026) — the next read on pricing and margin: a guide that holds the ~81% gross margin and sequential price gains pushes toward the bull case; the first sign of moderation confirms the base.
  2. DRAM/NAND contract-pricing prints — monthly spot and contract data into the first half of 2026 are the highest-frequency tell on whether the spike is normalising; the mechanism that ends every memory up-cycle.
  3. Industry capex announcements — FY27 capital plans from Micron, Hynix and Samsung; the supply ordered at the top is what builds the next trough.
  4. Hyperscaler / OEM inventory commentary — any sign of inventory rebuild or elasticity in PCs and phones would confirm demand is being pulled forward by price.

Faircurve Equities · Independent Single-Name Research · MU · 1 Jun 2026 · Issue No. 016

Sources: Financial Modeling Prep (quotes, statements, key metrics, analyst estimates & price targets, transcripts, price history). 8 quarterly earnings call transcripts (Q3 FY24 through Q2 FY26). Micron Technology SEC filings (10-K FY25, 10-Q Q2 FY26). Korean peer caps converted at ₩1,410/$.

Methodology: Forward FY+1 peer comparison across the memory & storage complex. Because the subject is a commodity cyclical at a margin peak, the base values normalised mid-cycle earnings power (not the peak run-rate) at a mid-cycle multiple. 40/40/20 blend of Forward P/E, Forward EV/Sales (substituted for EV/EBITDA where peers screen n/m) and a cost-of-capital DCF (13.8% discount rate, 3.0% terminal growth), cross-checked against a reverse-DCF of what the price implies.

Disclaimer: Research, not investment advice. Author has no position in MU at publication.