When: $197-200 with RSI oversold reset AND reversal candle on volume

Then: Add on confirmation

When: Close above $221.39 (10 EMA) on >150M volume with positive MACD histogram

Then: Accelerated deployment

Reduce position size by 10-15% ahead of FOMC; maintain position only with disciplined stop at 197, widening stops for beta-adjusted volatility around the meeting, and put hedge covering at least 50% of the position

• Working capital deterioration: Q2 operating cash flow halved from $50.34B to $24.08B QoQ; inventory grew 111% faster than revenue; receivables grew 127% faster than revenue
• Five bearish technical signals converging: loss of 10 EMA, MACD bearish cross with negative histogram, RSI neutral without support, close below middle Bollinger band, below VWMA for two consecutive sessions
• High beta (2.22) makes 5% downside a best-case scenario; hawkish Fed could produce 10% drawdown to test 50 SMA at 212.36 en route to 197
• Risk-reward math is closer to 1:1 than 1:3: realistic 10% downside vs 10-13% upside requiring reclaiming multiple resistance levels amid hawkish FOMC
• Polymarket pricing 82% probability of September hike and 90% probability of any 2026 hike; Barron's and institutional narrative shifting against AI durability
• DeepSeek and 160,000 Huawei Ascend cluster signal hyperscalers building alternatives to NVIDIA; one customer reducing forward orders could compress pricing power

→ vs aggressive: Working capital buildup is not evidence of demand strength; it's mathematical fact that inventory and receivables grew faster than revenue, historically preceding demand inflection

→ vs aggressive: Technical breakdown is not normal 5% pullback; five independent bearish signals converging in three days indicates distribution, not noise

→ vs aggressive: Macro dismissal is dangerous; when 10-year pushes toward 5%, customer multiples compress, capex budgets get revisited, and growth concentration risk gets reined in

→ vs aggressive: AI bubble defense using 2010 cloud/2016 smartphone analogy ignores that AI capex is measured against revenue expectations requiring flawless execution; Barron's is an early warning system, not background noise

→ vs aggressive: Risk-reward of 1:3/1:4 is not honest; realistic scenario is 10% expected downside vs 10-13% upside with low probability path through multiple resistance levels

HOLD with modifications: maintain core position, size put hedge at 25-30% of position (not 50%), require RSI reset and reversal candle confirmation before adding at $197-$212 zones, keep $197 hard stop as absolute line, and trigger re-evaluation after September FOMC based on actual outcomes

• NVIDIA fundamentals are extraordinary with Forward PE of 14, PEG of 0.55, ROE of 117%, and 75% gross margins
• Q2'27 FCF of $21.40 billion on $96.22 billion revenue still represents exceptional FCF margin despite drop from $48.59 billion
• TTM FCF of $41.81 billion on $5.27 trillion market cap yields approximately 0.79% FCF
• Receivables grew 127% versus revenue 106%—a 21 percentage point gap suggesting payment term changes
• Primary trend remains bullish with 50 SMA at 212.36 intact and 200 SMA at 197 intact and rising
• Current setup is a 5% pullback within uptrend, not distribution—50 SMA and 200 SMA both holding
• Technical setup shows path of least resistance sideways to down over 1-2 weeks: lost 10 EMA, MACD bearish cross, below VWMA for two sessions
• Macro risk is real but magnitude overweighted by conservative analyst—a 25bp hike compresses multiples 5-10%, not 20%
• Polymarket showing 82% probability for September hike and 90% for any 2026 hike
• Beta of 2.22 means 7-10% drawdown possible very quickly, especially into hawkish Fed
• Hedge sizing at 50% of position is cost-prohibitive given elevated IV on $5.27 trillion market cap name
• December $200-$210 puts covering 25-30% of position is appropriate—structured as collar to reduce cost
• Staged add Zone 1: $207-$212 requires RSI bottoming and volume declining confirmation
• Staged add Zone 2: $197-$200 requires 200 SMA holding plus RSI toward oversold and reversal candle on volume
• $197 hard stop is the structural floor anchor—tighten mental stops intraweek but keep hard stop absolute
• Portfolio construction check: if NVDA exceeds 15-20% of equity book, concentration warrants review
• FOMC scenario analysis: clean 25bp hike likely rallies on relief; hawkish surprise tests 50 SMA at 212; Fed pause validates aggressive thesis
• Hold captures fundamental upside without fighting technical reality, respects macro without capitulating
• Three modifications to HOLD: (1) tighten hedge to 25-30%, (2) require RSI reset and reversal candle before adds, (3) re-evaluate after September FOMC based on actual outcome

→ vs aggressive: Hand-waving working capital build as benign 'Blackwell stockpiling' without interrogating the 21 percentage point receivables-to-revenue gap

→ vs aggressive: Technical dismissal too clean—five indicators flashing caution in three days is not noise even if 200 SMA at 197 holds

→ vs aggressive: Macro dismissal most vulnerable—hyperscaler capex is not rate-insensitive; multiple compression on NVDA as high-beta proxy is real

→ vs aggressive: AI bubble dismissal intellectually lazy—secular thesis being right doesn't prevent 50% equity drawdown in multiple compression (Cisco 80 to 8)

→ vs conservative: Working capital point valid but overweighted—FCF of $21.40 billion is still massive in absolute terms and reflects strategic inventory positioning

→ vs conservative: Technical distribution call too confident—this is a pullback within uptrend, not regime change; primary trend remains bullish

→ vs conservative: Macro direction right but magnitude overweighted—5-10% drawdown over a week or two, not 20% collapse over three months

→ vs conservative: AI bubble concern real but overweighted—Barron's piece is one article; institutional narrative shifts take 6-12 months

→ vs conservative: 50% hedge sizing too aggressive and cost-prohibitive at elevated IV on $5.27 trillion market cap name