When: Price pulls back to $440-455

Then: Reload with stop at $430 (below July 29 capitulation low of $429.56)

When: Volume-confirmed close >$527 on ≥25M shares + 50 SMA holding as support + MACD line cross positive + RSI >55 sustained 3+ sessions, OR successful 50 SMA retest hold with higher-low formation

Then: Upgrade to Overweight

When: All Overweight triggers PLUS Q3 2026 print showing FCF recovery with sequential inventory normalization and forward 2027 EPS revisions upward

Then: Upgrade to Buy

When: MI450 delivers all five bull-case data points: named hyperscaler orders, perf/W benchmarks closing Nvidia gap, software ecosystem maturity, GM guide-up, ASP commentary

Then: Upgrade on confirmation

When: MI450 is 'good but not great'

Then: Wait for post-event fade to play out; reload at $440-455

HOLD with tighter framework: maintain current exposure but do not add on first test of $527 resistance; if AMD closes above $527 on 30 million shares for two consecutive sessions and pulls back to hold above $520, add a half-sized position with a stop at $505 (not $510); target weight 2-3% of portfolio (not 3-4%); trim above 4%; consider protective puts at $480 strike for 60 days; if bear trigger fires below $480, wait for weekly close below $440 with declining volume before considering reload.

• AMD ran from $171 to $584 (3.4x move) under hawkish macro, meaning easy money already made and stock is more vulnerable to mean reversion
• PEG of 0.49 relies on consensus Forward EPS of $15.63 that assumes flawless MI450 execution, gross margins at 53-54%, hyperscaler spending acceleration, and no Nvidia pricing pressure — any miss on any input compresses the multiple
• Beta of 2.48 means S&P drop of 5% implies AMD drops 12-13%; a 10% S&P correction could send AMD to test 200 SMA at $345, erasing the entire run from $440
• Sep 9-10 showed distribution: stock pushed to $526.79 on 22.2 million shares, reversed to close at $503.60 on declining volume (-4.4% intraday fade) — this is selling-the-news behavior, not healthy digestion
• AMD's AI growth is dependent on Microsoft, Meta, Oracle, and a handful of hyperscalers — sophisticated buyers with changing roadmaps who could shift capex to Nvidia's software ecosystem or better TCO; binary exposure dressed as secular growth
• Q2 2026: operating cash flow $2.37 billion but CapEx nearly tripled YoY to $808 million; FCF declined -39% QoQ; FCF yield is only 1.1%
• Revenue growth decelerating from 50% YoY to 30% as law of large numbers hits $11.5 billion quarterly base; CapEx becomes drag on returns and potential stranded capacity if MI450 disappoints
• Coiled spring breakout setups resolve in both directions — mid-June 2025 saw identical setup (fresh MACD crossover, neutral RSI, base breakout) before AMD fell 27.5% from $584 to $424
• A close below $480 opens path to $440, then $424, then $345 — 30%+ drawdown from current levels
• ARK funds explicitly trade around product cycles with proprietary customer pipeline visibility — their AMD trim ahead of MI450 is data, not noise
• AMD is currently 1.3% above 50 SMA and bumping into clearly defined resistance at $526 that has been tested and rejected once already
• Buy-in at $527 on first test of resistance risks being caught in failed breakout with stops at $510 that cascade into forced sales at $480/$440 if support breaks

→ vs aggressive: Dismissed hawkish macro as 'old news' citing AMD's 3.4x run from $171 to $584 — this is recency bias; past tailwinds don't justify forward aggression and parabolic rises make stocks MORE vulnerable to mean reversion, not less

→ vs aggressive: Called ARK trim 'portfolio mechanics' — but ARK's flagship funds explicitly publish research on trading around product cycles; their trim is informed positioning from a thematic fund with proprietary pipeline visibility, not random rebalancing

→ vs aggressive: Framed PEG of 0.49 as half of fair value without disclosing that Forward EPS of $15.63 is a consensus estimate that assumes flawless execution across multiple binary inputs; if any assumption breaks, 32x multiple becomes 50x on a smaller number

→ vs aggressive: Treated beta of 2.48 as a footnote when it's a 'flashing red light' — in hawkish macro with multiple catalysts for yield/spread stress, AMD moves 2.5x the market, making the 'asymmetric' setup only asymmetric if macro stays benign

→ vs aggressive: Described Sep 9-10 volume decline as 'healthy digestion' — intellectually dishonest framing of distribution behavior; failed test of $527 with declining volume is exactly the setup that precedes 15-20% corrections in extended names

→ vs aggressive: Dismissed customer concentration risk quickly — AMD's narrative is dependent on 3-4 hyperscalers who could redirect 2027 capex to Nvidia's mature software ecosystem or better TCO; this is binary exposure being mispriced as secular growth

→ vs aggressive: Celebrated revenue acceleration while ignoring FCF decline of -39% QoQ; FCF yield of 1.1% in a heavy investment phase means the best quarter in AMD's history also signals the easy comps are behind us

→ vs aggressive: Bull reload plan at $440-455 after $480 break misses the risk — $440-455 is still inside the prior consolidation and the next leg down targets $424 then $345; wait for weekly close below $440 with declining volume to confirm capitulation

HOLD with modifications: tightened bull confirmation (close above 527 on 25M+ shares), volatility-adjusted sizing (trim if AMD position dollar volatility exceeds 1.5% of portfolio), collar structure instead of naked puts (sell call at 560+ to finance put at 460 or lower), bear reload on close below 480 on heavy volume with stop at 430.

• Hawkish-and-accelerating macro (PPI re-accelerating, yields breaking higher, ECB hiking, oil spiking) is different from hawkish-but-stable that AMD ran through in 2025; macro inputs are actively worsening rather than stable
• Technical signals (MACD cross Sep 8, RSI at 55 with room to 70, ATR compression from 40 to 22) are observable facts about supply and demand that should not be dismissed
• Valuation is conditional on execution: Forward EPS of $15.63 embeds assumptions about MI450 success, gross margin continuation, and AI capex acceleration
• Sep 9-10 reversal (4.4% intraday fade on declining volume after high-volume push to new highs) is failed breakout behavior; has implications for how aggressive to be on retest
• AMD at 3-4% weight behaves like 7-8% position in a low-beta name during stress events due to beta of 2.48
• AMD sizing depends on portfolio context: tech-heavy portfolios with NVDA, AVGO, MU, SMCI have high correlation with AI infrastructure drawdowns, making AMD effectively larger
• Bear reload zone of 440-455 is well-placed; buying at bottom of prior consolidation after failed breakout is historically good entry in primary uptrends
• Both aggressive and conservative analysts optimize for one scenario; analyst is positioning for both with probability-weighted outcome of grind higher with elevated volatility

→ vs aggressive: Macro dismissiveness is wrong: hawkish-and-accelerating (PPI re-accelerating, ECB hiking, oil spiking on geopolitical event) is different from hawkish-but-stable; if Fed tightens further, same technical setup breaks the other way

→ vs aggressive: 0.49 PEG is not a free option; Forward EPS of $15.63 embeds assumptions that could be cut, rapidly compressing the PEG

→ vs aggressive: ARK has been wrong on timing consistently (trimmed Tesla at $200 multiple times, missed Nvidia's biggest moves by being too cautious on catalyst window); treating single trim as gospel is incorrect

→ vs aggressive: Buying the breakout zone on first test without confirmation is buying the most dangerous level on the chart

→ vs aggressive: Collar structure caps upside, which aggressive analyst will dislike, but that is correct reaction and why it is right structural hedge

→ vs conservative: Dismissing bullish technical setup carries zero information is not risk management, it's paranoid positioning; MACD, RSI, and ATR signals are real observable facts

→ vs conservative: Long puts at 480 strike for 60 days is expensive given implied volatility; cost drags on returns if AMD doesn't break down

→ vs conservative: Waiting for weekly close below 440 with declining volume before reloading is too patient; close below 480 on heavy volume is a tradable signal