Inference Unit Economics in the 2025 Cycle
How throughput, context windows, and GPU memory tiers shift margins across model‑hosting vendors.
Investment Notes
Selected research memos and commentary. Subscribe for updates or reach out for LP‑only materials.
Signals from Enterprise AI Pilots to Production
Tracking conversion rates, feature retention, and budget pathways across the Fortune 500.
Platform Effects in the AI Tooling Layer
Assessing defensibility: distribution vs. differentiation in developer platforms.
Position Sizing with Liquidity Guardrails
A practical framework for sizing, stop‑losses, and post‑mortems in concentrated portfolios.