When: Close <$430 (200 SMA)

Then: Break multi-quarter uptrend; re-engage full Sell review

HOLD and wait for pullback to $494-$499 entry zone; if price breaks above $522 on heavy volume upgrade conviction; if price breaks below $499 on rising volume stand aside entirely

• RSI bearish divergence confirmed: 73.53 today versus 79.15 on August 10 with price at 513.53 versus early August high around 505
• MACD histogram printing negative bars for seven consecutive sessions with expanding magnitude
• ATR contraction from 17 down to 12 indicating coiled volatility
• Bollinger Band: close at 513.53 sits above upper band at 511.58 by less than two points; intraday high was 517.78 — exhaustion at first test of resistance near $522 invalidation level
• Mean-reversion target at middle band 492.75 converging with 10 EMA at 494.66
• 10 EMA filled on August 17 at 479.45 with RSI cooling to 62 — mean reversion happened
• Position sizing: 3 to 5 percent; $34 risk per share; 20 to 35 basis points of portfolio risk on aggressive entry at $513
• Conservative entry plan at $494 to $499 with stop at $465 yields approximately 170 basis points maximum loss vs aggressive plan's approximately 175 basis points with higher probability of stop being hit
• Polymarket pricing Anthropic at 99 percent probability of frontier leadership — consensus extreme implies asymmetry to downside
• Polymarket shows 1 percent chance of September cut and 10 percent by December — asymmetric risk if hawkish AI capex phrasing triggers multiple compression
• 3 to 5 percent intraday drawdown on Jackson Hole day is not unusual for stocks at 52-week highs
• Valuation at 22 times forward earnings with 31 percent expected EPS growth requires perfect execution on $116 billion capex translation, $73 billion deferred revenue backlog conversion, and benign macro regime
• Trailing free cash flow of $16.55 billion versus historical run rates of $60 to $70 billion — FCF 75 percent below historical norms
• Operating cash flow of $183 billion covers capital return program of $47 billion annual dividend and buyback, but buffer narrows quickly if AI demand softens 10 to 15 percent
• Stock trading 19 percent above 50-day moving average with momentum indicators rolling over on binary Fed event day
• $331 billion in revenue, 45 percent operating margins, $73 billion deferred revenue backlog — all real but insufficient as short-term entry signal

→ vs aggressive: RSI divergence is real not 'lagging indicator noise' — price making higher high while momentum fails to confirm at 73.53 versus 79.15; MACD histogram printing seven consecutive negative bars with expanding magnitude while price grinds higher on thinning participation

→ vs aggressive: Bollinger Band close barely above upper band at 511.58 with intraday high at 517.78 testing near $522 invalidation level is exhaustion, not 'textbook breakout signature' — mean-reversion target at middle band 492.75 and 10 EMA at 494.66 is high-probability support

→ vs aggressive: 'Pullback may never fill' is dangerous narrative — 10 EMA filled on August 17 at 479.45; patience trade at $494-$499 entry gives 14 additional points of cushion versus chasing $513 with same $34 stop distance, improving probability-weighted expected value

→ vs aggressive: OpenAI plus Anthropic at 70 percent of AI revenue is demand risk not a toll booth moat — Polymarket pricing Anthropic at 99 percent probability of frontier leadership is consensus extreme implying asymmetry to downside; single disappointing model release, governance event, or regulatory action would compress multiple faster than re-rate higher

→ vs aggressive: Jackson Hole binary event has asymmetric second-derivative risk regardless of 'priced in' probabilities — hawkish AI capex phrasing would hit MSFT multiple specifically because entire bull case rests on sustained AI demand at scale

→ vs aggressive: 22 times forward earnings with 31 percent EPS growth prices in the bull case — trailing FCF of $16.55 billion versus historical $60 to $70 billion is market signaling it doesn't fully believe capex translation story yet

→ vs aggressive: FCF compression has real consequences — institutional holders running dividend-focused or income-overlay strategies cannot hold names where FCF is 75 percent below historical norms; $47 billion annual buyback is flexible capital allocation, not permanent floor, and gets reduced first if demand softens 10 to 15 percent

HOLD with triggered scaled entry—starter position of 1.5 to 2% at current levels with hard stop at 479; add remaining 1.5 to 3% only on pullback confirmation (492-500 zone with volume and MACD divergence) or breakout confirmation (above 522 on heavy volume with positive MACD)

• MACD histogram negative for seven consecutive sessions while price makes new highs signals weakening tape, not just noise
• ATR contracted from 17 to 12 indicating volatility compression and easy buying exhaustion
• Close at 513.53 barely two points above upper band at 511.58; intraday high at 517.78 rejected below 522 invalidation level
• 10 EMA rising at roughly 5 points per day—conservative analyst's 494-499 fill zone is a moving target shrinking daily
• Concentration risk is real but overstated; $73 billion deferred revenue backlog and long-term contracts create substantial switching costs
• Jackson Hole positioning already prices in hawkish outcome; trader should wait until Monday to add regardless of speech result
• Starter position of 1.5-2% with 34-point stop risks only 13 basis points of portfolio—appropriate given stretched technicals
• Both breakout above 522 and successful retest of 492-500 are valid confirmation triggers for adds

→ vs aggressive: Dismissal of MACD deterioration is 'intellectually lazy'—seven consecutive negative histogram bars while price grinds higher reflects narrative-supported price action, not healthy momentum

→ vs aggressive: Bollinger band argument needs work—close at 513.53 barely above upper band at 511.58 and rejection at 517.78 below 522 invalidation is buyer exhaustion, not continuation signature

→ vs conservative: Wait for 494-499 plan has fatal flaw—10 EMA rising at 5 points per day means the fill zone shrinks every day and may never materialize on exact terms

→ vs conservative: Jackson Hole argument is weakest part of case—market already positioned for hawkish outcome; the plan creates a coin-flip between missing a breakout or getting a pullback that isn't asymmetric

→ vs conservative: Concentration risk overstated—Microsoft is platform layer with enormous switching costs; governance event at OpenAI doesn't mean Anthropic migrates to AWS