When: Preference to avoid outright trim

Then: Implement 60-90 day collar (sell $220 call / buy $180 put) to define binary risk without crystallizing floor

HOLD with trim guidance: reduce 15-25% above $200, do not add above $190, establish starter positions only on pullback to $170-175 with stabilizing FCF and inventory normalization. Stop at $160.

• TTM FCF is $10.2 billion, but Q2 FY26 FCF of $495 million was second consecutive quarter of sequential decline
• Q2 inventory build of nearly $1 billion consumed cash; management issued $1.99 billion in new short-term debt to fund buybacks
• Inventory grew 32% YoY while revenue declined 4% YoY; no corroborating book-to-bill commentary or channel checks
• 9/18 candle was wide-range bearish engulfing on 54.4 million shares (3x average daily volume), not absorption
• Stock closed at $194.23, one penny below upper Bollinger band at $194.33 — exhaustion risk elevated
• ATR is 8.92; 50 SMA at $167.94; 200 SMA at $167.13; medium-term trend has not confirmed short-term move
• Forward P/E of 19x with PEG of 0.79 anchored to consensus EPS of $10.20, which can change if December disappoints
• Data center pivot supported only by one reported Amazon framework (undisclosed minimums/margins) and optical interconnect demo — not revenue or purchase orders
• 52-week low is $121.99; 52-week high is $259.92
• Bear case: compression to 50 SMA is ~14% drawdown; revisit 200 SMA or worse is 15-30% downside
• Bull case requires multiple things to go right simultaneously; bear case requires only one thing to go wrong — asymmetric risk in wrong direction
• Fed on hold with 96% probability of no cuts; market pricing soft landing with 6% recession odds — no macro tailwind left
• Stock up 37% over six months at upper Bollinger band into binary catalyst
• Stop at $160 protects against December disappointment and re-rating toward 200 SMA

→ vs aggressive: FCF yield as floor ignores that company is borrowing $1.99 billion to fund capital returns — not seasonal noise, yellow flag becoming red if Q3 prints weak

→ vs aggressive: Inventory +32% framing as 'smart over-ordering' lacks corroborating evidence; inventory growing while revenue contracting suggests uncertainty, not deliberate pre-launch build

→ vs aggressive: 9/18 'textbook absorption' misreads the tape — 54.4M shares on bearish engulfing is distribution, not absorption; short-covering tactical buyers, not confirmed long-term accumulation

→ vs aggressive: Valuation comparison to AMD 2017-2018 EPYC pivot is apples to oranges; AMD had proven silicon, third-party benchmarks, and revenue-generating roadmap at re-rate — QCOM has a demo and a framework

→ vs aggressive: Asymmetry argument works against bulls — requires multiple things to go right; bear case only requires one miss on data center ramp

→ vs aggressive: Position sizing advice (full size into December) is precisely how career-ending drawdowns happen; original trim plan (15-25% above $200, no new longs above $190) preserves capital through binary events

HOLD with modified position sizing: maintain core position, trim ~10% only above $200 with volume confirmation, do not add above $190, tighten stop to $170 (not $160), wait until after December quarter prints before sizing up

• AMD 2017-2018 comparison is flawed — QCOM situation fundamentally different with no merchant CPU revenue, no third-party benchmark wins, and only a reported Amazon framework (not a design win or purchase order)
• Amazon framework is not just a demo — reported $60 billion purchasing framework represents meaningful hyperscaler validation that didn't exist six months ago
• Q2 FCF annualization to $2 billion is wrong, but TTM FCF of $10.2 billion ignores the downward trajectory — truth is in the middle
• $1.99 billion new short-term debt issued in Q2 to fund buybacks is a concern if Q3 prints weak — sustainable for one quarter but not a pattern
• 9/18 candle was massive volume at lows with powerful reversal next session — real pattern but doesn't invalidate 54-million-share distribution event
• Contested tape at upper Bollinger band after 7% headline-driven move — acknowledge bid but don't assume it persists without confirmation
• Bull case requires multiple things to go right simultaneously; bear case only requires one to break — asymmetry is real but magnitude of Amazon framework shouldn't be dismissed
• VWMA at $179.65 and 10 EMA at $170 are the real structural support levels, not the round number $160
• December quarter will not resolve cleanly in either direction — market will spend 4-6 weeks pricing implications before actual print
• Snapdragon Summit this week is a real-time signal for whether to add on pullback or stand aside

→ vs aggressive: AMD 2017-2018 comparison is intellectually lazy — QCOM lacks merchant CPU revenue, third-party benchmark wins, and hyperscaler design wins that were generating dollars; calling a trade show demo a design win is being early and hoping the market validates timing, not analytical rigor

→ vs aggressive: Treating the one-day 9/18 bounce as structural support when it may just be short-covering into a headline catalyst is unsound — a contested tape at upper Bollinger after a 7% move requires acknowledgment of the bid but not assumption it will persist

→ vs conservative: Amazon framework dismissal ignores the magnitude of what happened — a $60 billion purchasing framework represents hyperscaler validation regardless of undisclosed minimums and margins; market is forward-looking and the market is telling you something

→ vs conservative: Position sizing guidance of trimming 15-25% above $200 and waiting for $170-175 pullback forces selling into momentum, waiting for a move that may never come if catalyst resolves debate, and then chasing back in at higher prices after confirmation — that is a directional bet framed as prudence

→ vs conservative: Overweighting the bear scenario ignores the magnitude of the Amazon framework attachment; being 'conservative' with position sizing is itself a directional decision, not a neutral one