Bear wins the central analytical debate. Forward PE of 5x is the most damning signal — the market is already pricing in dramatic earnings normalization, and in memory specifically, the market has historically been right about cycle peaks. ROE of 88% has no historical analogue in DRAM — Samsung topped at 45%, Hynix at 40%, even Micron's 1999-2000 peak didn't sustain 80%+ — and the "this time is different because HBM" thesis is exactly the narrative that has preceded every prior memory cycle crash. Technicals confirm deterioration: failed breakout 12.5% below prior high, MACD making four consecutive lower highs (36→35→33→30), RSI declining from 66 to 54, two closes below the 10 EMA, and below-average volume on the post-earnings drift. The macro backdrop (10Y at 2002 highs, 96% probability of zero Fed cuts) is unsupportive of a 2.2-beta $1.18T name. The receivables jump (5.6x vs. revenue 4.4x) signals expanding DSO — a classic late-cycle quality flag. Where credit is given: HBM does have more long-term contracted volume than prior DRAM cycles, the U.S.-supplier geopolitical premium is real, and the $29B FCF/net-cash position eliminates financial distress risk — which is why not going full Sell. The asymmetry favors caution: the bull case requires ALL of (HBM ASPs holding, hyperscaler capex holding, no new entrants qualifying, margins sustaining >50%, multiple expansion) — the bear case requires only ONE to break. This is a 1-3 quarter tactical underweight, not a structural call against MU as a company — the HBM structural argument has merit for the 2027+ timeframe once memory pricing resets.

POSITION STRATEGY: Trim existing MU exposure by 30-50% to reduce high-beta (2.2) exposure into deteriorating technicals and a hostile macro backdrop.

INITIATION RESTRICTION: Do not initiate new long positions at current levels — the risk/reward is asymmetric to the downside given the 5x forward PE signals and failed breakout pattern.

STOP MANAGEMENT: Bear invalidation is a daily close above $1,100 (10 EMA reclaim + prior high retest at $1,097 with conviction) — if that fails to materialize, maintain underweight bias.

SCALING PLAN: Consider adding to short/underweight on any rally toward $1,097-$1,100 (failed-breakout retest zone) rather than chasing weakness, to maximize risk/reward.

HEDGE COVER LOGIC: Cover any short hedge if MU closes above $1,100 with volume AND MACD turns up — this would invalidate the failed-breakout thesis.

WATCH LIST: Watch list for adding back: (a) multiple compression plays out with forward PE expanding back toward 10-12x as the market prices normalization, (b) HBM ASP commentary softens confirming cycle peak, (c) technical basing pattern emerges near $850-$913 (BB lower / July low retest).

PAIR TRADE: For accounts that benchmark to tech/semis, consider pairing the underweight with continued overweight in lower-beta AI beneficiaries (e.g., select foundries or equipment names with beta < 1.5) to maintain AI exposure with less drawdown risk.

TIME HORIZON: This is a 1-3 quarter tactical underweight, not a structural call against MU as a company — the HBM structural argument has merit for the 2027+ timeframe once memory pricing resets.

The research plan's central thesis is that MU is at or near a memory cycle peak, evidenced by an unsustainably low forward PE of 5x, an 88% ROE with no historical DRAM analogue (vs. Samsung's 45%, Hynix's 40%), and deteriorating technicals including a failed breakout, four consecutive MACD lower highs (36→35→33→30), declining RSI (66→54), and closes below the 10 EMA on below-average volume — classic distribution signatures. The bear case requires only one of (HBM ASPs, hyperscaler capex, no new entrants, margins >50%, multiple expansion) to break, while the bull case requires all five to hold — the asymmetry decisively favors caution. I am expressing the underweight view by scaling into selling on a rally toward the failed-breakout retest zone, which maximizes risk/reward versus chasing weakness.

Stop loss: $1100.00

Sizing: Trim 30-50% of existing MU exposure; for new underweight adds, size to 1-2% of portfolio notional given MU's 2.2 beta and $1.18T market cap sensitivity to macro/tech rotation.