SELL

• Technical picture is broken: failed twice at 1011, MACD rolled over from 11.70 to 8.62 in three days, RSI non-confirmation at 59.5, price below 10 EMA (937) and 50 SMA (963), VWAP rollover at 924, closed at 910
• Catalyst calendar is hostile: NVDA earnings Wednesday with 97% beat priced in, Jackson Hole Friday with hawkish Warsh (Polymarket pricing no-cuts narrative at 86%), YMTC IPO injecting $4.9 billion into Chinese memory competitors
• Traders plan to sell into 950-1011 zone assumes a courtesy rally that broken momentum may not deliver; either sell now at 910 or use tighter limits at 940-950
• Left tail risk extends to 780 or worse: ATR at 67 points, Bollinger width at 26% of price, single-session moves of 100-130 points possible on hawkish surprise or NVDA guide-down
• Conservative analyst recommends trimming 50-60% of overweight (not 25-40%), leaving 30% of position exposed vs 70% under current plan
• 80% operating margin is a cyclical peak, not a moat; memory is a commodity business with ASP normalization risk
• $155 forward EPS could fall to $90-100 on 30% ASP normalization, turning 5.87 forward P/E into 9-10
• Dividend yield is only 0.05%; there is no income cushion to offset volatility
• Beta at 2.213 means full-size hold through catalyst cluster has deeply negative risk-adjusted return
• Re-entry conditions: daily close above 963 with rising volume and improving forward estimates, not just a close above 1011
• Hard stops: cut further if closed below 880; exit entirely if closed below 766 (breaks 766-1027 range, reopening path toward 200 SMA at 574)

→ vs aggressive: Only trimming 30% leaves 70% of position exposed to the full left tail; that's not a hedge, it's a token gesture

→ vs aggressive: The asymmetry framing is incomplete; worst-case isn't 850 but 780 or worse given 67-point ATR and single-session move potential of 100-130 points

→ vs neutral: Using long-term 200 SMA (574) to justify inaction ignores short-term risk window between now and Friday's catalysts

→ vs neutral: Waiting for confirmation often arrives after the worst damage has been done; holding has opportunity cost, volatility cost, and drawdown cost

→ vs neutral: $22 billion in customer commitments through 2030 doesn't protect against industry-wide supply expansion; YMTC is raising $4.9 billion to do exactly that

SELL/TRIM MU for existing holders (25%–40% staged reduction); HOLD/CASH for accounts without position

• Reduce 25%–40% of existing MU position through staged limits rather than full liquidation
• Staged approach: sell one-third near current market, second tranche if bounce to $937–$950 fails, final tranche near failed $963–$1,011 retest
• Price rejected twice near $1,011, below 10-day EMA ($937) and 50-day SMA ($963)
• MACD histogram still positive, 200-day SMA still rising, close at $910.43 near Bollinger midpoint (~$897)
• Higher lows from July low near $737 indicate damaged but not confirmed reversal
• Balance sheet strengths: net cash, 3.43 current ratio, lower debt, customer commitments through 2030
• Counterweights: receivables surge, elevated capex, potential Chinese capacity build (YMTC IPO)
• Beta of 2.213 indicates elevated market sensitivity but does not guarantee proportional moves
• Forward EPS $155 (extraordinary assumptions); P/E at $910 = 5.87x; if EPS falls to $90, multiple = ~10x; at $120, ~7.6x
• ATR of $67; $766 area is lower Bollinger and prior-support zone, not automatic target
• If close below $880 with expanding volume, cut another 10%–15% of original position
• If close below ~$766 with estimate cuts and weaker memory pricing, eliminate tactical sleeve, limit strategic position to 10–25 bps
• First re-add after close above ~$963 with volume improvement and stable-to-higher estimates
• Second re-add after close above $1,011 and successful retest of that level
• Hold proceeds in cash or broad-market fund rather than another semiconductor/AI name for genuine diversification

→ vs aggressive: Calling this a confirmed distribution phase or clear short setup is too certain; close at $910 was near equilibrium (~$14 above Bollinger midpoint ~$897), not in confirmed breakdown; MACD line above signal and histogram positive; rising 200-day SMA providing long-term support

→ vs aggressive: Beta of 2.213 indicates elevated historical market sensitivity but does not mean 5% market decline produces exact 11% MU decline

→ vs aggressive: 97% Nvidia beat probability is not 97% probability of positive MU guidance; upside surprise remains possible

→ vs aggressive: 'Unlimited downside' is incorrect; claimed 4% portfolio hedge overstated—trimming 30% of MU position and avoiding $40 decline protects ~1.2% of full position value, not 4% unless MU is nearly entire portfolio

→ vs aggressive: Two $1,011 rejections are concerning but not automatically textbook higher-high failure; second test does not clearly establish new higher high

→ vs aggressive: One wide-range down day with ~30M shares does not prove institutional distribution, especially alongside sector-wide weakness

→ vs aggressive: Historical memory-cycle comparisons are not evidence by themselves; 80% operating margin, 346% revenue growth, $155 EPS are extraordinary assumptions that could prove wrong

→ vs aggressive: Would not initiate naked short solely because MACD histogram is shrinking

→ vs conservative: 50%–60% cut applies same certainty in opposite direction as aggressive view; demand and financial data remain unusually strong despite sector-wide de-risking

→ vs conservative: 86% probability of no 2026 rate cuts from prediction market, not verified policy forecast; should not be treated as exact risk measure

→ vs conservative: $4.9B YMTC IPO is structural overhang, not $4.9B of immediately available competing DRAM/HBM capacity

→ vs conservative: Hard stops presented as likely base-case; $766 area is important support zone, not proof MU is headed directly to 200-day SMA

→ vs conservative: Daily close below $880 is warning but not automatic exit signal with $67 ATR

→ vs conservative: Execution problems: $980 is limit-order reference, not expected entry price; 'stop loss' at $1,011 is above current price and cannot function as protective stop for long position

→ vs conservative: Ambiguous phrase 'reduce overweight by 25%–40%'—percentage of overweight not same as percentage of total holding; plan must specify starting weight, target weight, and portfolio value

→ vs conservative: 25–50 basis-point cap is upper limit on residual tactical exposure, not actual hedge; 50-bps position can still experience 20% drawdown creating ~10-bps portfolio impact

→ vs conservative: Waiting for rally to $980 or $1,011 may mean missing intended exit given major catalysts ahead

→ vs conservative: Selling MU and buying another semiconductor or AI-related stock provides no genuine diversification given same technology and memory-cycle risk

→ vs conservative: Full strategic exit only appropriate if 6–18 month thesis invalidated; single technical break should not automatically force complete liquidation