Memory & AI Compute

This strategy invests in a curated basket of U.S. companies powering the AI revolution — think memory chipmakers, semiconductor equipment giants, and the networking and compute firms that keep AI data centers running. It's long-only, meaning it profits when these companies grow, and it never bets against the market. Rides the explosive demand for AI training and inference hardware — from HBM memory makers to GPU architects and the fabs that produce them. As AI model sizes double every 6 months, memory bandwidth and compute density are the binding constraints. This strategy targets the companies whose order books are already locked in for 2-3 years of capacity.