CoreWeave Lambda Crusoe Top Three

This ranking shows that cloud GPU revenue is concentrating around three very different operating models, one built for hyperscaler scale, one built for developer flexibility, and one built for giant data center projects. CoreWeave is the clear leader at $3.52B in trailing revenue as of June 2025, Lambda is the next largest at about $505M annualized as of March 2025, and Crusoe follows at $276M in 2024 revenue, which makes the top three less a single market than three separate ways to package scarce Nvidia capacity.

• CoreWeave is much larger because it signs multi year contracts with buyers like Microsoft and scales whole facilities around those commitments. That produces more revenue and higher gross margin, but it also pulls the business toward a small set of very large customers and giant capex cycles.
  1 sacra 3 sacra
• Lambda sits in the middle by selling owned GPU capacity in much smaller chunks. Its software can carve a 16,000 plus GPU cluster into short duration slices, so research teams and startups can rent dedicated infrastructure without committing to thousands of GPUs for years at a time.
  1 sacra 2 sacra
• Crusoe rounds out the top three with a more infrastructure heavy model tied to large buildouts like Oracle and Stargate. That makes it a major GPU cloud by revenue, but one whose growth is driven more by financing and campus development than by the self serve developer motion that defines Lambda.
  1 sacra 4 sacra

The next phase is likely to widen the gap between these three leaders and the rest of the field. As more capital flows into data centers, power, and long term chip supply, CoreWeave should deepen its hyperscaler position, Lambda should keep winning flexible developer workloads, and Crusoe should convert megaproject construction into cloud revenue at industrial scale.