Summary of Earnings (My View

Key Highlights:
-    Revenue: $46.74B – up 56% year-over-year, comfortably above consensus (~$46.05B).
-    Adjusted EPS: $1.05, beating expectations (~$1.01).
-    Data Center Revenue: $41.1B (up 56% YoY), but just shy of forecast (~$41.3B).
-    Stock Buyback: $60B approved—on top of $24.3B in H1 returns via buybacks/dividends.
-    Q3 Revenue Guidance: Projected at $54B (±2%), slightly above Street ~ $53–$53.8B.
-    Post-Market Reaction: Shares slipped 2–3% despite strong metrics—reflecting investor caution.

CEO Jensen's Earnings Call Highlights:
Blackwell AI Chips Demand
– "Everything's sold out" Demand for H100s, H200s, and the Blackwell line remains extremely strong.
– Blackwell Ultra is ramping at full speed, hailed as “the AI platform the world has been waiting for.”
China Opportunity Still Critical
– Jensen reiterated that China represents a potentially $50B market, stressing its significance (second-largest computing market; home to ~50% of AI researchers)
– Ongoing discussions with the U.S. government: Nvidia seeks greater access to China, despited the 15% revenue-sharing terms tied to the H20 chip export licenses.
Product Roadmap & AI Evolution
– Rubin chips teaser: slated for next year, maintaining Nvidia’s annual product release cadence.
– Jensen emphasized the shift toward “reasoning AI” and physical AI, which demand vastly more compute but offer dramatically improved accuracy ("hallucinations have dropped significantly").
AI Startup Boom
– AI-native ventures revenue could surge to $200B next year from ~$20B this year. A 10× expansion, fueled by open-source models and widespread adoption.
H20 Chip Status
– No H20 shipments to China yet; however, some $180M in excess inventory was sold elsewhere.
– CFO estimates $2–$5B in potential H20 revenue for Q3 if geopolitical conditions permit.

What I would do:
Immediately Post-earnings: Hold current exposure; avoid chasing upward. Deploy up to 10% of fund into defined-risk option spreads (e.g., vertical spreads) to hedge against potential pullback or volatility.
Short-Term (1–2 weeks): Monitor developments regarding China H20 access and demand signals from hyperscalers. Consider adding 5–10% more exposure if positive clarity emerges.
Mid-Term: If Rubin roadmap gains traction and China access improves, consider scaling NVDA position to 15–20%; otherwise, reallocate to emerging AI names or defensive sectors.
Risk Control: Set alert zones: trim or hedge if NVDA dips >5–7% post-earnings, or cut exposure if data center trends show sustained weakness.