Patience, conviction, and the limits of being early.
Dear partners,
Q2 closed up 6.2% against SPY, with the bulk of attribution from NVDA and the quiet networking sleeve. We trimmed PLTR on the late-April rip and added selectively into the May drawdown. The book is now 18% cash — the highest in twelve months — not because we are bearish, but because optionality is unusually cheap right now and we like having dry powder when narrative shifts.
The harder lesson this quarter was a smaller one: a watchlist name we'd held conviction on for two years finally re-rated, and we hadn't sized it. Being early is being wrong, as the saying goes; the corollary is that being right at the wrong size is also a form of being wrong.
We continue to underwrite the agent-infrastructure thesis (picks & shovels in the AI capex cycle) and remain net long the index. Specific names and rationale follow in the position notes below.
— A.N.
This is a parody letter. None of this is investment advice. The numbers, like the LPs, are imaginary in proportion.