CoreWeave (CRWV) investor relations material

Business performance and demand environment

• Achieved $5.1 billion in FY2025 revenue, with nearly $100 billion in backlog and $40 billion in new bookings in a single quarter.

• Demand for AI infrastructure is described as insatiable, with customers being turned away due to capacity constraints.

• Client base spans hyperscalers, AI labs (including nine of the top 10 global non-China labs), and rapidly growing enterprise customers.

• Enterprise segment is expanding, with a record number of new enterprise logos added in Q4.

• Financial services represent over $10 billion of the current backlog, heavily utilizing infrastructure for inference workloads.

Technology and product strategy

• Focused on building and operating large-scale GPU infrastructure, primarily using Nvidia GPUs, with seamless support for both training and inference workloads.

• Software stack (Mission Control, SUNK) enables efficient provisioning and operation of clusters at scale.

• Infrastructure is designed for flexibility, allowing rapid transition between training and inference, often within hours.

• Clients are not demanding edge or inference-only facilities, nor alternative chip architectures; demand remains centered on Nvidia GPUs.

• Older GPU generations (Ampere, Tesla, Volta) remain profitable and in use, with six-year depreciation cycles considered conservative.

Financial structure and contract dynamics

• Contracts typically span four to six years, with 25% contribution margin targets across all client segments.

• Long-term take-or-pay agreements are used to de-risk capital investments and ensure high utilization rates.

• Clients often expand contracts as their usage grows, with renewals and expansions preferred over short-term on-demand pools.

• Margin compression in Q1 is expected to reverse as new deployments ramp in Q2 and Q3, supporting guidance for low double-digit operating margins by year-end and a long-term target of 25%-30%.