When CoreWeave went public earlier this year, we laid out a clear and ambitious roadmap. We said we would scale faster, diversify our customer base, expand our cloud platform, and remain at the forefront of AI as workloads move from experimentation to production.

I am immensely proud of how we delivered on that plan in 2025.

Along the way, AI entered a new phase. Demand accelerated rapidly, expectations rose just as fast, and infrastructure shifted from a consideration to a defining constraint. AI labs had proven what was possible. The next challenge became enabling the broader ecosystem to build with AI and adopt it at scale. In that environment, our focus was simple: execute with discipline and build the Essential Cloud for AI.

Going public was about committing to invest for the long-term while operating with rigor in a market highly constrained by the availability of compute, power, and infrastructure. Since the IPO, we scaled capacity aggressively, broadened our critical services through a combination of organic development and strategic acquisitions, strengthened our balance sheet to drive down the cost of capital and fuel growth, and deepened relationships with the world’s most demanding AI builders. That foundation allowed us to move up and down the technology stack as AI continued to proliferate beyond the frontier. 

The result is a business that is broader, more resilient, and better positioned than ever.

We also said we would lead on performance and reliability, and we did. Independent industry analysts repeatedly recognized CoreWeave as the leading AI cloud for performance, scalability, and reliability. That consistency matters. It reflects years of thoughtful engineering across our entire technology stack, and it is why customers trust CoreWeave to power their most critical AI workloads.

To meet relentless demand we scaled globally, growing from roughly 900 employees to more than 2,000 worldwide. Since the beginning of the year, we brought 9 new data centers online across the United States and Europe for a total of 42 active data centers in our network. We also broke ground on many more, including our first greenfield build in our home state of New Jersey, a flagship facility in Pennsylvania, and multiple sites in the United Kingdom. By the end of Q3, we reached approximately 590 megawatts of active power, with a clear path toward exceeding 850 megawatts by year’s end across a diverse portfolio of 2.9GW of contracted capacity. 

CoreWeave became an end-to-end platform for AI builders. We delivered industry-leading AI object storage, launched the first publicly available serverless reinforcement learning product, and continued to advance CoreWeave Mission Control, the operating standard for all AI use cases. Strategic acquisitions from Weights & Biases and Marimo to  Monolith and OpenPipe accelerated our evolution and added new team members critical to our continued  growth. 

Put simply, CoreWeave has built, and continues to build, a tightly integrated suite of infrastructure and software that enables customers to build with and run AI systems continuously, reliably, and at global scale.

As a result of this strong foundation we diversified our customer base materially, demonstrating that CoreWeave’s platform is built for flexibility across industries, use cases, and generations of technology. AI labs and startups continued to rely on our technologies, while we expanded further to power large enterprises and public-sector workloads. We launched CoreWeave Ventures to support founders shaping the AI ecosystem and extended our platform to begin meeting the security and performance requirements of public sector organizations with CoreWeave Federal.

All of this progress unfolded  as the world experienced extraordinary pressure to deliver on AI’s promise, from supply-chain constraints to heightened expectations around reliability, efficiency, and responsible deployment. In this environment, CoreWeave’s people ran toward the hardest problems and delivered, setting the standard through consistency, rigor, and ownership every day.

This is what it means to be the Essential Cloud for AI, and it is the work we are doubling down on in 2026. Thank you to our teams for their extraordinary execution, and to our customers and partners for their trust and collaboration, who made this progress possible.

We said we would build for this moment. We did. More to come in 2026 and beyond.