The author predicts that MU, META, and NVDA will outperform NBIS, IREN, and CRWV over the next 12 months. The thesis suggests that "Neoclouds" are becoming commodities and will face multiple compression as hyperscalers build their own compute solutions. Consequently, the market opportunity for assets like NBIS is expected to decline significantly from current levels.