Position Status: +15.9% YTD | Entry: $181.74 (2026-04-09) | Current: $210.73 | Conviction: 5→3/10 (deteriorating) --- 1. IS THE ORIGINAL THESIS STILL INTACT? Short answer: No. The thesis is fractured — and insider activity combined with macro shifts have exposed structural weaknesses I initially underweighted. My April thesis stood on three pillars: - Pricing power in AI infrastructure - High ROIC sustaining competitive moat - Margin resilience through cycles Here's what the evidence actually says now. What's Supporting the Thesis (Weakly) NVDA's asset base expanded from $206.8B (Jan 2026) to $259.5B (Apr 2026) — a 25% jump — which *could* signal healthy balance sheet absorption of earnings. But this is precisely the kind of superficial metric that flatters deteriorating fundamentals. Large cash positions inflate asset bases. They don't prove capital efficiency is improving. What's Actively Undermining It (Heavily) Insider selling at scale is the clearest red flag. - Mark Stevens (CFO) sold 500,000 shares on Jun 4 across two tranches (400k + 100k) - Stephen Neal sold 15,500 shares on Jun 3 - This is CFO-level liquidation, not routine diversification. The CFO doesn't unload half a million shares because he's rebalancing a portfolio. The competitive threat is real and accelerating. - Amazon is "in talks to sell custom AI chips" designed to erode NVDA's dominance - Custom silicon is a direct assault on NVIDIA's pricing power — the core of my original thesis - Meta, Google, Tesla are all building competing chips. That wasn't true in April. The macro backdrop has shifted the calculus. - Warsh's "higher-for-longer" indefinitely crushes re-rating theses for capital-intensive names - NVDA *is* capital-intensive. Higher rates compress the NPV of future AI capex cycles — full stop - The bond market's disagreement with Fed messaging (mortgage rates falling while the Fed holds firm) adds uncertainty, not tailwind Earnings quality concerns are mounting. - I don't have the detailed accrual breakdown from the latest 10-Q to assess cash conversion rigorously — but the velocity of insider selling at $210 tells its own story - If the CFO believed the AI infrastructure capex cycle had three or more years of real runway left, he wouldn't be dumping this aggressively at +16% YTD --- 2. WHAT'S SURPRISED ME ABOUT THIS POSITION'S PERFORMANCE? Two things. **First:

Research these companies