CEO Michael Intrator's 2025 Letter to Shareholders

CoreWeave 2025 Annual Report Shareholder Letter

Announcement

Announcement

Webinar

Announcement

Podcast

Announcement

GTC 2026

Announcement

CoreWeave brings up the industry’s first NVIDIA Vera Rubin NVL72 deployment.

Read more

Products

Data and storage

Infrastructure control

Runtime acceleration

Model and agent development

Mission control

Solutions

Pricing

Resources

About us

Contact us Login

Contact us Login

Clear

To our shareholders There are moments in history when a new technology stops being a curiosity and becomes the foundation for everything that follows, when even the fiercest skeptics become believers. We are in one of those moments. Artificial intelligence is no longer a bet on the future; it is the present. Those who recognize this shift are not waiting. They are building. What's remarkable is who is building: a new generation of pioneers. Their story follows a telling arc, one that began in universities, then frontier labs, then startups, and now enterprises across every industry. All of them seeking the pace, performance, and partnership required to drive the next wave of invention. CoreWeave counts itself among them. Our story began in a New Jersey garage before AI became inevitable. We saw early that the traditional approach to compute would be a bottleneck to progress. AI depends on massive parallel computation at a scale general purpose clouds are not designed to support. So we made a deliberate choice. We started with a clean slate and redefined a cloud built for this moment. What followed was audacious and defiant: a full-stack AI cloud, conceived when others doubted the opportunity and, frankly, us. That mindset became our strength, and with it the instinct to run with our customers toward the hardest problems. Most did not see us as an obvious choice to reinvent the cloud. Doubt, it turns out, is a powerful motivator. After our first calendar year as a public company, that conviction has translated into CoreWeave becoming the fastest cloud platform in history to reach $5 billion in annual revenue, while growing 168% year-over-year. Our cloud footprint now spans more than 850MW of active power across 43 data centers globally. Today, nine of the leading ten model providers rely on CoreWeave Cloud. AI has moved from possibility to prerequisite and we are proud to be at the center of this shift as The Essential Cloud for AI™. ‍ Discipline to build and scale Before we were cloud builders, we were risk managers. That instinct to size opportunities carefully, build against what we can see, and never outrun our contracts shaped how we financed CoreWeave from the beginning. It struck some as unconventional for the enterprise technology space. But many of the greatest infrastructure builds in history have been financed this way. And for good reason: it works. Our 2025 results are a direct reflection of that approach. We diversified our customer base significantly, growing CoreWeave Cloud customers 1 by approximately 150%, with no single customer representing more than 35% of our revenue backlog (vs. 85% to start the year) That backlog expanded to $66.8 billion, up from $15 billion, with weighted average contract length 2 extending from four years to five. We secured approximately $18 billion of debt and equity capital while reducing our weighted average cost of debt by more than 300 basis points. That translated into roughly $700 million in annualized savings based on our year-end debt balance.

We’ve shown the industry how this historic buildout can be financed, tying capital deployment directly to long-term contracted demand, and creating a new asset class in the process. The market is recognizing the quality of CoreWeave's financing model and track record of execution, reflected by the reduction of our cost of capital and confidence that AI adoption is accelerating. We intend to continue to innovate in the capital markets, unlocking new sources of financing at lower costs while partnering with many of the world's leading financial institutions. The strong financial foundation we have established is what allows us to keep building at the scale and speed this moment requires.

A deliberate choice to redefine cloud for this moment.

The differentiation of our platform runs deep, and it’s rooted in the belief that an AI cloud works best when technology, tools, and teams are designed as one. This year we scaled our infrastructure and software services, expanded our platform through organic innovation and strategic acquisitions, and invested in the talent and partnerships to deliver it. Unconstrained by tradition, we have a structural advantage that is both intentional and durable. The CoreWeave Cloud spans the full stack, from silicon and infrastructure to orchestration, tooling, and application-layer services. It encompasses GPUs and CPUs, purpose-built networking, storage architected for AI throughput, security isolation, and a support model designed to operate as a true technical partner. Because CoreWeave is trusted by the most sophisticated builders who sit closest to the toughest challenges, we have a singular sightline into what they need and we innovate from there. We focus on providing the most deeply differentiated technical solutions to the most ambitious customers on the planet. That specialization at scale is central to everything we do. To move at the pace our customers demand, we completed targeted acquisitions, each adding critical capabilities to accelerate our roadmap. With the addition of Weights & Biases, we brought model development, evaluation, and monitoring natively onto our platform. Together, our teams launched a new inference service and a set of tools for building and operating agentic applications. We also expanded into reinforcement learning through the acquisition of OpenPipe, making RL more accessible. And we introduced CoreWeave ARENA™, a lab where customers can benchmark real-world performance before moving workloads into production. Underpinning it all is CoreWeave Mission Control™. More than a product, it has become a preferred operating standard for how AI infrastructure should run. It brings together security, observability, and operational expertise to enable the most demanding workloads, maximizing performance and extending the useful life of infrastructure over time. The industry continues to recognize these advancements. We are consistently the first to bring next-generation hardware to…