CoreWeave (CRWV) UBS’s 2025 Global Technology and AI Conference summary

Demand trends and drivers

• Revenue growth reached 134% last quarter, with a $55 billion backlog, indicating robust and accelerating demand.

• Demand has experienced multiple step-function increases this year, driven by both existing and new customers.

• Compute needs are rising due to scaling laws, more intensive post-training, and inference workloads, especially with new AI models.

• Frontier AI labs are the primary demand drivers, but AI-native startups and traditional enterprises are increasingly contributing.

• The application layer for enterprise AI adoption is still in early stages, with efforts to help mature industries adopt AI through acquisitions like Monolith.

Supply chain and infrastructure scaling

• Building gigawatt-scale AI campuses is unprecedented, with ongoing supply chain challenges such as labor shortages and long lead-time equipment.

• Delays in project timelines have occurred, but most delayed capacity is expected to come online in Q1.

• Self-build projects and expanded data center teams are helping improve control and transparency over site development.

Market dynamics and risk perspectives

• Current infrastructure build-out is not speculative; capacity is being built to meet signed, long-term customer contracts.

• Demand is described as insatiable, with hyperscalers and providers struggling to keep up, countering concerns of an AI bubble.

• Pricing for GPU instances remains stable, with high utilization and customers willing to recontract at similar rates due to strong ROI.