LG Doucet and analyst Melvin examine why neoclouds such as CoreWeave and Nebius could capture a meaningful share of AI infrastructure spending. Unlike general-purpose cloud platforms, these companies optimize GPU capacity for model training and inference, while contracted backlogs and rising rental prices point to persistent demand.
The discussion compares CoreWeave’s scale and expanding backlog with Nebius’s faster growth, customer prepayments, and premium valuation. It also addresses the sector’s central risks—heavy capital spending, interest costs, power constraints, and unprofitable operations—and explains why SpaceX may become a formidable competitor without invalidating the broader growth thesis.