Solo PM. Small partner circle. No LP redemption gates, no monthly NAV optics, no risk-committee stop-outs, no benchmark-hugging pressure, no compliance-driven caps beyond our own.
We are not Citadel. We do not have 1,000 PMs, sub-second execution, satellite imagery, or tick-level alt-data. We will lose every contest where those things matter.
Short-leash mandates cannot hold a thesis through a multi-quarter re-rating, a guide cut, or a sentiment trough.
We can underwrite a multi-year thesis and tolerate interim drawdowns if the structural read is unchanged. A guide cut is a tax on short-leash books, not on us.
No P&L smoothing. Some quarters will be ugly. Drawdown rules are guards, not invalidation triggers.
Rigid single-name caps force diversification across thousands of books — it is enforced, not chosen.
Meaningful core positions in names we have actually done the work on. Concentration is the engine, not a side effect.
Idiosyncratic blow-up risk. The defense is depth of work, not breadth of holdings.
Pods need narratives that fit a one-page tearsheet. Messy capital structures, M&A integration noise, and segment restatements get under-owned.
We do the slow work — long filings, conference-call nuance, segment-level accounting. Slower coverage. Deeper ownership.
We pick fewer names and own them deeper. Coverage breadth is the trade.
Sell-side consensus is the gravity well. Most pods take the consensus number, sleeve it slightly, and trade the catalyst.
We bracket Goldman vs UBS vs Bernstein, place our number inside the bracket, and size against the disagreement.
Being early often looks like being wrong. We will be marked wrong before we are marked right.
Sector pods are siloed. Semis pod doesn't talk to software pod doesn't talk to China pod.
One brain owns the AI compute stack end-to-end, and reads Chinese-language sell-side next to western desks every morning.
No sector specialization moat. We have to earn cross-stack credibility every quarter.
Earnings ingestion, transcript Q&A pull, sell-side reconciliation, cover-memo drafting, watchlist scorecard grading — all run by agents on a deterministic workflow.
Most analysts read about agents. We run them. We feel the cost curve, the latency floor, the context-engineering tax. Operator-grade tape reading no spectator can fake.
We invest beneficiary-side, not producer-side. The product is not the AI; it is what the AI lets the operator do that they couldn't before.
