When: Sustained Google Cloud growth >50% with disclosed operating margin in Q3/Q4 2026 prints

Then: Add back to position

When: Free cash flow turning positive again in Q3 or Q4 2026 prints despite continued capex

Then: Add back to position

When: Alphabet discloses material external TPU revenue contribution

Then: Add back to position

When: Meta's consumer AI agent push shows no measurable search share impact in advertising data

Then: Add back to position

HOLD existing positions, do not add capital, tighten risk controls — wait for price confirmation above 200 SMA on rising volume with MACD histogram contracting before committing fresh capital

• 200 SMA at $335.96 is acting as overhead resistance, not support — price closed at $330.65, below all four key averages (10 EMA, 50 SMA, 200 SMA, 20 VWMA)
• MACD widening negative at -3.58 with accelerating histogram; RSI grinding toward oversold at 39.47 with no confirmed bullish divergence
• Q2 2026 FCF of -$5.9 billion driven by $39.1B operating cash flow against $44.9B CapEx; buybacks went to zero
• Total long-term debt quadrupled year-over-year from $23.6B to $98.2B; $24.8B in new debt issued Q2; preferred stock line emerged at $18B from zero
• Clean normalized forward EPS around $14.85 versus $19.93 trailing — headline GAAP distorted by $98.8B gain on security sales
• Forward PE of 22.3x is the real valuation, not the misleading 16.6x trailing PE
• Prediction market odds of no Fed cuts in 2026 at 93%, rising 4.1 percentage points in one week — repricing to more restrictive stance, not priced in
• Meta re-rated +6.1% on competitive dynamics in consumer AI, threatening Alphabet's 60% margin search bar business
• Realistic downside if 200 SMA breaks is $273-$280 (15-17%), not the 6% computed from $314 stop placed using 2x ATR

→ vs aggressive: Selling at 200 SMA is not panic-selling — the 200 SMA at $335.96 is currently overhead resistance with price at $330.65; below the 200 SMA it flips from floor to ceiling

→ vs aggressive: Negative FCF is not a bullish signal — it's a yellow flag with flashing lights; the company spent $5.9B more than it generated and issued $24.8B in new debt while buybacks went to zero

→ vs aggressive: The $314 stop is anchored on 2x ATR from entry, but ATR measures recent volatility not stress volatility; if 200 SMA fails, the next reference is $273-$280, not $314

→ vs aggressive: Morgan Stanley put Meta in pole position with +6.1% re-rating in consumer AI — this threatens Alphabet's 60% margin search business, not just a nibble at the edges

→ vs aggressive: 93% odds of no Fed cuts in 2026 rising 4.1 percentage points is repricing to restrictive, not pricing in — long-duration assets like GOOGL face compression

→ vs aggressive: 16.6x trailing PE is a mirage; clean forward PE of 22.3x is the real valuation, and at that multiple with capital discipline deteriorating, you're paying full price

HOLD existing position; do not add fresh capital; refine stop to tiered structure; consider 25% partial hedge/collar; define explicit re-engagement triggers

• 6% downside risk-reward math is flawed as it assumes stop at $314 holds; in stressed mega-cap breakdown, next 5% down often happens in a week and stops get run 2-4% through
• Real probability-weighted downside in breakdown scenario is $273-$280 zone, not $314 stop level — 15-17% from current, not 6%
• Confirmation defined as close above 200 SMA at $335.96 with rising volume and contracting MACD histogram; without this, position is in confirmed downtrend with a story, not uptrend with a dip
• Meta AI agent value chain shift risk could compress Alphabet's search business from 60% to 25% margin even without Alphabet 'losing' — structural risk worth monitoring
• 200 SMA loss is only 1.58% — one closed session below level that was held for weeks; full breakdown not yet confirmed
• Balance sheet remains strong: $242B cash and short-term investments, $42B net debt, $922B asset base, $640B equity — not a leveraged story despite negative -$5.9B Q2 FCF vs $39B operating CF
• Stop placement suggestion: tiered structure with primary invalidation at lower Bollinger Band ~$325 on closing basis, hard risk exit only on weekly close below $314
• Hedge suggestion: take 25% of position and roll into partial collar or trim, converting binary position to asymmetric (capped downside, retained 75% upside)
• Re-engagement triggers: (1) GOOGL closes above $335.96 on volume 10%+ above 20-day average with MACD histogram stopping expansion — consider 30-50% scale-in; (2) Cloud 82% growth confirmed/beats next print — add trigger regardless of price action
• If neither trigger fires in 60-90 days, re-evaluate entire thesis — catalysts aren't translating
• Cost of waiting for confirmation is missing first 3-5% of move if it works; cost of not waiting is catching falling knife in sector with 4 consecutive down sessions and widening MACD

→ vs aggressive: Risk-reward framing is seductive but wrong — 6% downside assumes stop at $314 holds, but correlated selling in mega-cap tech means next 5% down happens in a week, not a month, and stops get run 2-4% through in real breakdown; true probability-weighted downside is $273-$280, not $314

→ vs aggressive: Confirmation is not a theoretical gate — it has specific, observable definition (close above $335.96 with rising volume and contracting MACD); without it, trader is in confirmed downtrend with a story, not uptrend with a dip

→ vs conservative: Bear case framing is too clean — 200 SMA was lost by only 1.58% in one session, not confirmed by weeks of breakdown action; assuming worst-case on first day after level gives way is as anchoring as aggressive analyst's assumptions

→ vs conservative: Balance sheet framing mischaracterizes situation — $242B cash, $42B net debt on $922B asset base is not leveraged; -$5.9B Q2 FCF was one quarter against $39B operating CF; the issue is ROIC on capex, not solvency

→ vs conservative: Confirmation framework has no terminal state — if stock grinds sideways for weeks and stages slow rally on fundamentals without ever producing clean reclaim, trader stays in cash indefinitely; that's process paralysis, not discipline