CoreWeave
| Company type | Public |
|---|---|
| Industry | Cloud computing, Artificial intelligence |
| Predecessor | Atlantic Crypto |
| Founded | 1 January 2017 in Livingston, New Jersey, U.S. |
| Founder | Michael Intrator, Brian Venturo, Brannin McBee, Peter Salanki |
| Headquarters | Livingston , United States |
Area served | United States, Europe |
Key people | Michael Intrator (CEO), Brian Venturo (CSO), Brannin McBee (CDO), Peter Salanki (CTO), Nitin Agrawal (CFO), Sachin Jain (COO), Chen Goldberg (SVP Engineering), Michelle O’Rourke (Chief People Officer), Glenn Hutchins (Lead Independent Director) |
| Products | Cloud-hosted GPU infrastructure, proprietary chip management software, Mission Control software, CoreWeave Cloud Platform |
Number of employees | 1450 (2025) |
| Website | coreweave |
CoreWeave, Inc. is a U.S.-based company specializing in artificial intelligence (AI) cloud computing, with its headquarters located in Livingston, New Jersey. The firm focuses on delivering cloud-hosted graphics processing unit (GPU) infrastructure tailored for AI developers and corporate clients, alongside creating its own proprietary chip management software.
Established in 2017 under the name Atlantic Crypto, initially concentrating on high-performance computing, CoreWeave now manages data centers throughout the United States and Europe. It supports a diverse client base, including both multiple tenants and exclusive single customers. Notably, the company operates a $1.6 billion supercomputer facility for Nvidia in Plano, Texas, which Nvidia acknowledges as the world’s fastest AI supercomputer.
2017–2021
CoreWeave was founded in 2017 in New Jersey by commodities traders Michael Intrator, Brian Venturo, Brannin McBee, and Peter Salanki. Originally named Atlantic Crypto, the business focused on mining ethereum cryptocurrency using graphics processing units (GPUs). After the downturn in cryptocurrency markets during 2018, the company rebranded to CoreWeave in 2019 and shifted its business model to provide cloud computing services by utilizing its substantial GPU assets.
2022–2023
By 2022, CoreWeave had expanded to operate three data centers within the United States. During that summer, it committed roughly $100 million to acquire Nvidia’s latest H100 GPUs. In October 2022, CoreWeave introduced an accelerator initiative designed to support startups with compute credits, discounts, and access to hardware resources on its cloud platform.
Responding to the surge in AI processing needs throughout 2022 and 2023, CoreWeave leveraged its privileged access to Nvidia GPUs and secured clients such as Stability AI. The company also enhanced its infrastructure by deploying extensive fiber-optic networks. In April 2023, Nvidia invested $100 million into CoreWeave, which reached an estimated valuation near $2 billion by May.
In August 2023, CoreWeave arranged a groundbreaking $2.3 billion debt financing deal led by Magnetar Capital and Blackstone. This transaction was notable for using Nvidia’s H100 GPUs as collateral—the first instance of such hardware being employed in this manner—drawing widespread attention within financial media.
2024
In May 2024, CoreWeave secured $1.1 billion in funding with leadership from Coatue Management, elevating its valuation to $19 billion. By October of the same year, an investment from Cisco was anticipated to boost the valuation further to approximately $23 billion. Concurrently, CoreWeave obtained a $650 million credit facility from leading Wall Street banks including Goldman Sachs, JPMorgan Chase, and Morgan Stanley to support expansion efforts across its operations and data centers.
The company announced plans to allocate $978.6 million toward developing data centers in the U.K. during 2024. In June 2024, CoreWeave made a $1 billion acquisition proposal for HPC provider Core Scientific, which was ultimately declined. By November 2024, CoreWeave became one of the inaugural cloud service providers to integrate Nvidia's new Blackwell GPU technology.
Also in November, CoreWeave completed a $650 million secondary share offering involving investors such as Jane Street Capital, Magnetar Capital, Fidelity Management, and Macquarie Capital. Reports from Bloomberg indicated that CoreWeave planned an initial public offering (IPO) for 2025, appointing Morgan Stanley, Goldman Sachs, and JP Morgan Chase as lead underwriters. The company filed its registration statement (Form S-1) in March 2025 with intentions to list on the Nasdaq exchange under the ticker symbol "CRWV". The filing revealed that over 60 percent of CoreWeave’s projected revenue for 2024 would come from business with Microsoft.
2025
In February 2025, CoreWeave distinguished itself as the first cloud service provider to deliver Nvidia GB200 NVL72 chips through cloud computing. IBM revealed intentions to utilize these GB200 clusters for training its Granite AI model. The following month, CoreWeave acquired Weights & Biases, an AI platform developer, for roughly $1.7 billion. March also saw OpenAI enter into a five-year, $12 billion cloud computing agreement with CoreWeave, which included a $350 million private placement during CoreWeave’s IPO, granting OpenAI an ownership interest.
On March 27, CoreWeave scaled back its IPO target from $2.7 billion to $1.5 billion and went public the next day, successfully raising $1.5 billion. This offering marked the largest AI-focused IPO by capital raised at that time.
By July 2025, CoreWeave became the pioneer in commercial deployment of Nvidia Blackwell Ultra GPUs (GB300 NVL72), utilizing servers supplied by Dell.
CoreWeave reached an agreement in July 2025 to acquire Core Scientific for $9 billion; however, this acquisition was declined by Core Scientific shareholders in October.
In September 2025, CoreWeave announced plans to purchase OpenPipe, a startup specializing in reinforcement learning tools designed for training AI agents. The subsequent month, it agreed to acquire Monolith AI, a company developing AI and machine learning applications centered on physics, aiming to broaden its industrial cloud service portfolio.
Key people and employees
Based in New Jersey, CoreWeave employed approximately 1,450 staff members as of 2025. Michael Intrator holds the position of CEO and founder. The company's cofounders include Brian Venturo as chief strategy officer, Brannin McBee as chief development officer, and Peter Salanki serving as CTO. In 2024, Nitin Agrawal joined from Google to take on the CFO role. That same year, Sachin Jain—formerly with Oracle’s AI division—was appointed COO in August, alongside Chen Goldberg from Google who became senior vice president of engineering. Michelle O’Rourke was appointed chief people officer in October 2024, while Glenn Hutchins assumed the role of lead independent director in February 2025.
Cloud platform and software
CoreWeave’s Cloud Platform is constructed upon a Kubernetes-native framework optimized for extensive GPU-intensive workloads. Designed specifically for AI development, it caters to developers creating AI-driven solutions. The company’s Mission Control software provides clients with tools to monitor and validate hardware performance, thereby improving oversight and management of their cloud environments.
Data centers
CoreWeave operates multiple data centers across the United States and Europe. By 2025, the company expanded its network to 32 data centers housing 250,000 GPUs—a significant growth from 15 centers in 2024. In 2024, CoreWeave inaugurated its London headquarters along with two UK-based data centers. Additionally, that year it secured a lease for a $1.2 billion facility covering 280,000 square feet at the Northeast Science and Technology Center located in Kenilworth, New Jersey.
Some facilities are designed to serve numerous clients simultaneously, whereas others are dedicated exclusively to individual customers or specialized projects. A notable example is the $1.6 billion supercomputing complex in Plano, Texas, constructed for Nvidia; spanning 450,000 square feet and containing over 3,500 H100 GPUs, this installation was recognized as the world’s fastest AI supercomputer in 2023.
CoreWeave also leases high-performance computing resources from Core Scientific. In December 2024, Canadian AI startup Cohere announced plans for a multi-billion-dollar data center developed and managed by CoreWeave. By January 2025, two UK data centers equipped with Nvidia Hopper (H200) GPUs became operational following their development throughout 2024. In that year, 77% of CoreWeave’s revenue originated from its two largest clients, with Microsoft alone contributing 62%.
In December 2024, CoreWeave collaborated with Dell Technologies on infrastructure projects. During 2025, CoreWeave and Microsoft jointly invested in establishing AI Hub in New Jersey—a partnership involving the New Jersey Economic Development Authority and Princeton University.