CoreWeave, Inc. is an American cloud company specializing in workloads that require substantial computing power. Based in Livingston, New Jersey, it provides graphics processing units, or GPUs, along with the software tools, storage and networking needed to use them. Its positioning differs from that of general-purpose cloud providers: it offers infrastructure designed around accelerated computing, which is particularly sought after for artificial intelligence.
From cryptocurrency mining to cloud computing for AI
Founded in 2017 by Michael Intrator, Brian Venturo and Brannin McBee, CoreWeave has its roots in cryptocurrency mining, particularly ether. This activity gave it experience in operating GPUs at scale. The company then redirected its resources toward providing on-demand computing for applications such as graphics rendering and, later, machine learning.
The rise of generative AI has accelerated this transformation. The training requirements of large models create significant demand for interconnected GPU clusters available over extended periods. CoreWeave is expanding its infrastructure and attracting investors, including Nvidia, which is also a key supplier of its accelerators. Its Nasdaq listing in March 2025, under the ticker symbol CRWV, marked a new stage in financing its development.
Infrastructure designed for intensive computing
The core offering is access to GPUs in data centers, accompanied by services for orchestrating workloads. CoreWeave relies notably on Kubernetes to manage containerized applications. The aim is to enable customers to deploy substantial resources without purchasing the equipment themselves or taking responsibility for all aspects of its physical operation.
Use cases include training artificial intelligence models, adapting them and performing inference, meaning running them once deployed. Scientific computing and visual rendering are also among the areas served. Beyond the chips themselves, performance depends on communication between machines, data access and the reliability of the system as a whole. CoreWeave brings these different building blocks together to serve technology companies, AI laboratories and organizations with intensive computing needs.
What next?
CoreWeave’s trajectory will depend on its ability to turn demand for AI into a sustainable business. Its expansion in the United States and Europe involves securing equipment, sites and an adequate power supply. It also requires substantial financing, as new generations of processors follow one another rapidly.
Facing major cloud providers and other GPU specialists, the company will need to maintain a balance between growth, infrastructure utilization rates and debt management. The diversification of its customer base and the development of inference will also be areas to watch. The challenge now goes beyond access to chips: it concerns the ability to deliver a reliable, economically sustainable service over the long term.