When: Anthropic S-1 confirmation with day-one valuation ≥$1.5T and October pricing

Then: Upgrade to Overweight

When: Q3 AWS reacceleration >18% YoY with FCF stabilization

Then: Upgrade to Overweight

When: 10Y yield breaks below 4.5% sustained

Then: Upgrade to Overweight

When: Price at $238-242 with RSI <35 AND MACD histogram contracting

Then: Staged add position

When: Confirmed close >$256 on >55M shares with MACD/RSI confirmation

Then: Staged add position

HOLD at 50% sizing with 240 stop; trim into strength at 254-258; do not widen stops to 235-237; wait for MACD/RSI confirmation before adding

• Forward P/E is 24x on a stock with declining forward EPS (-16%), not a generational entry at $287 breakout that has now failed
• Q2 2026 EPS of $5.75 includes $15.68B gain on sale of securities; clean EPS closer to $2.50-$3.00
• Anthropic $1.7B revenue (0.2% of revenue) signals multi-cloud optionality away from AWS exclusivity, not insignificance
• Free cash flow collapsed from $32.9B (2024) to $7.7B (2025) to negative $18.2B (Q1 2026)
• Long-term debt nearly doubled from ~$65B to $119B in one quarter ($54B increase); interest expense about to step up materially
• 10-year at 5%, Fed 96% probability of no cuts, biased toward hikes, rates at 2007 highs
• Technical stack bearish: below 10 EMA, 50 SMA, Bollinger middle band; MACD negative with expanding histogram; RSI mid-40s with bearish divergence at August highs
• Aggressive analyst's 240 stop is correctly placed at structural level; widening to 235-237 adds ~50% more dollar risk per share on a 1.44 beta name
• Three binary catalysts in October: Prime Day, AWS Q3 disclosure, FOMC
• Current 50% sizing protects capital; full sizing would turn a 9% drawdown into 15-20% drawdown

→ vs aggressive: Generational entry framing is a narrative trap; stock rallied to $287 three weeks ago, failed at breakout with bearish MACD divergence, now below 10 EMA, 50 SMA, and Bollinger middle band—not capitulation, just a failed breakout

→ vs aggressive: Forward P/E 24x on declining earnings requires flawless execution to defend; PEG of 1.48 is not cheap

→ vs aggressive: Anthropic 0.2% of revenue framing is mathematically true but strategically wrong; market prices AWS thesis durability, not individual line items

→ vs aggressive: Free cash flow collapsed from $32.9B to $7.7B to negative $18.2B; balance sheet leveraged at worst possible time with 10-year at 5%

→ vs aggressive: ATR compression resolution is a coin flip; technicals show distribution, not a coiled spring

→ vs aggressive: 235-237 stop is 13-15 dollar risk vs 9 dollar current stop; 50% more dollar risk per share for a 1.44 beta name in macro shock environment

→ vs aggressive: Full sizing into three binary October catalysts with 10-year at 5% and negative FCF is not conviction; current 50% sizing is asymmetric risk management

→ vs aggressive: Trimming at 254-258 is not selling the catalyst; it's harvesting premium at unconfirmed technical level that already failed once

HOLD AMZN at 50% sizing, $240 stop, trim 25-33% into $254-258 resistance; do not add until clean close above $256.18 with MACD crossing above signal, RSI above 50, and volume above VWMA at $252.38

• Rating: HOLD - maintain current position without aggressive adding or removing
• Position sizing: 50% of full conviction weight is appropriate given October binary catalysts (Prime Day, AWS Q3 disclosure, FOMC) and 10-year at 5% with hawkish Fed
• Stop placement: $240 at the 200 SMA structural line—widening to $235-237 absorbs 50% more dollar risk per share in a setup where the core thesis is invalidated
• Trim 25-33% into $254-258 resistance zone that has already failed once; trim is harvesting premium, not selling catalyst
• Do not add until full confirmation: clean close above 50 SMA ($256.18) with MACD crossing above signal, RSI above 50, volume above VWMA ($252.38)
• 200 SMA at $240.85 is rising and intact—price still above it; distribution not confirmed until this line breaks on daily close
• Anthropic-Akamai ($1.7B against $775B revenue) is minor narrative ding to AWS moat, not thesis break
• Forward P/E at 24x with forward EPS declining 16% is uncomfortable; optical Q2 EPS inflated by $15.68B non-recurring gain—normalized earnings closer to $2.50-3.00 vs $5.75
• Cash and short-term investments: $143 billion; debt-to-equity at 45.6% is elevated but not distressed; interest coverage remains strong
• Upgrade triggers: 10-year rolls below 4.5%, confirmed AWS re-acceleration in Q3, or clean reclaim of $256 with full momentum confirmation
• Downgrade triggers: 200 SMA breaks on daily close with volume, confirmed Fed hike, or meaningful Anthropic-tier customer defection broadening AWS narrative concern
• Stock is 27% above February low and failed breakout at $287—this is a pullback, not a crash; do not load boat on pullbacks

→ vs aggressive: Wrong to frame this as 'generational entry'—a generational entry requires something cut in half on heavy volume with narrative capitulation, not a pullback from failed breakout at $287 that remains 27% above February lows

→ vs aggressive: ATR compression does not signal direction—compressed ATR means volatility is coming, not that resolution will be higher; trends also break

→ vs aggressive: Full sizing request is backward-looking based on trailing earnings and ROE; October has Prime Day sell-the-news risk, AWS Q3 disclosure, and FOMC with hawkish Fed—this is when to be smaller, not bigger

→ vs aggressive: Should not widen stop to $235-237; this absorbs 50% more dollar risk per share to stay in a trade where the structural thesis has been invalidated; with 1.44 beta and hawkish macro, gap through 237 overnight is real risk

→ vs aggressive: Do not add at $256 reclaim without full momentum confirmation—a tactical bounce without MACD and RSI confirmation is not a confirmed trend resumption

→ vs conservative: Wrong to call this distribution—distribution requires price to close below 200 SMA first; until that happens, this is a pullback, not distribution

→ vs conservative: Balance sheet is not 'broken' or 'distressed'—$143B in cash, 45.6% debt-to-equity is elevated but not distressed, interest coverage remains strong; the company is investing aggressively in AI capex supercycle using debt as part of funding mix

→ vs conservative: Anthropic deal is not a thesis break on its own—the real question is whether Anthropic is the canary, not whether this one deal matters in isolation; both aggressive and conservative are using it as bigger piece of thesis than it deserves

→ vs conservative: Q2 EPS gains are not 'accounting fiction'—$15.68B gain on securities is real cash that came in the door; it's non-recurring, so the multiple should price off normalized earnings, but that is different from calling earnings fake