CoreWeave, Inc. is an American artificial intelligence (AI) cloud-computing company headquartered in Livingston, New Jersey. It provides cloud-based graphics processing unit (GPU) infrastructure to AI developers and enterprises, and also develops its own chip management software. Founded in 2017 as Atlantic Crypto, the company transitioned from cryptocurrency mining to cloud computing and has grown into a major provider of high-performance computing services, culminating in its initial public offering on the Nasdaq in March 2025.
The company operates its own data centers in the United States and Europe, with some facilities dedicated to multiple companies and others to a single client. Its $1.6 billion supercomputer data center for Nvidia in Plano, Texas, has been described by Nvidia as the fastest AI supercomputer in the world. CoreWeave's business model centers on providing access to Nvidia's latest GPU hardware, including the H100, GB200, and Blackwell Ultra chips, often before other cloud providers.
Founding and Early History (2017-2021)
CoreWeave was founded in 2017 in New Jersey by three commodities traders - Michael Intrator, Brian Venturo, and Brannin McBee - along with Peter Salanki. The company was originally known as Atlantic Crypto and initially operated as a cryptocurrency mining company, mining ethereum using graphics processing units. In the wake of the 2018 cryptocurrency crash, the company was renamed CoreWeave in 2019, leveraging its large inventory of GPUs to start providing cloud computing infrastructure to companies.
The pivot from mining to cloud services allowed CoreWeave to capitalize on its existing hardware investments. By 2021, the company had established a foothold in the emerging field of GPU-based cloud computing, though it remained a relatively small player compared to established cloud providers like Amazon Web Services, Microsoft Azure, and Google Cloud.
Growth and Nvidia Partnership (2022-2023)
CoreWeave was operating three data centers in 2022, all in the United States. In the summer of 2022, the company invested heavily in Nvidia's latest and fastest H100 chips, spending around $100 million on the purchases. This strategic bet proved prescient as demand for AI processing surged. CoreWeave launched an accelerator program in October 2022, which gives startup companies compute credits in addition to discounts and other hardware resources on the CoreWeave cloud.
As market demand for AI processing increased in 2022 and 2023, CoreWeave, which had unique access to Nvidia GPUs, saw its business jump considerably. The company signed clients such as Stability AI and continued to buy chips and build out new data centers, with its data center technicians installing 6,000 miles of fiber-optic cabling in 2023. In April 2023, Nvidia invested $100 million in CoreWeave, and the company was valued around $2 billion by May 2023.
In August 2023, CoreWeave secured a $2.3 billion debt financing facility led by Magnetar Capital and Blackstone by using Nvidia's H100 GPUs as collateral. The deal marked the first time that H100-based hardware had been used as collateral, which attracted some note in the business press. This innovative financing approach demonstrated the perceived value of the GPU hardware and CoreWeave's ability to leverage its assets.
Expansion and Valuation Increases (2024)
The company raised $1.1 billion in funding led by Coatue Management in May 2024, reportedly valuing the company at $19 billion. According to a Bloomberg News report in October 2024, Cisco was set to invest in CoreWeave as well, reportedly valuing the company at $23 billion. Also in October 2024, CoreWeave announced that it secured a $650 million credit line for expanding its operations and data centers, with Goldman Sachs, JPMorgan Chase, and Morgan Stanley among those who led the financing.
The company had announced a plan by 2024 to invest $978.6 million in U.K. data centers. In June 2024, CoreWeave's $1 billion offer to buy high-performance computing capacity vendor Core Scientific was rejected. In November 2024, Tom's Hardware reported that CoreWeave was among the first cloud service providers to receive shipments of Nvidia's new Blackwell hardware.
In November 2024, CoreWeave closed a $650 million secondary share sale, with investors such as Jane Street, Magnetar Capital, Fidelity Management, and Macquarie Capital. Bloomberg reported in November 2024 that CoreWeave was planning a 2025 initial public offering and had selected Morgan Stanley, Goldman Sachs, and JPMorgan Chase to manage the process.
Initial Public Offering (2025)
CoreWeave filed its Form S-1 in March 2025 and planned to list on the Nasdaq under the symbol "CRWV". According to the filing, Microsoft accounted for over 60 percent of CoreWeave's revenue in 2024. In February 2025, CoreWeave was reported to be the first cloud provider to make Nvidia GB200 NVL72 chips available via cloud computing, and IBM announced it would use the GB200 clusters to train its Granite AI.
In March 2025, CoreWeave announced the acquisition of the AI platform developer Weights & Biases, reportedly for around $1.7 billion. Also in March, OpenAI signed a five-year cloud-computing contract worth approximately $12 billion with CoreWeave for its AI infrastructure needs. The deal allowed OpenAI to acquire a stake in CoreWeave through a private placement of $350 million worth of shares during the IPO.
CoreWeave reduced its IPO size from $2.7 billion to $1.5 billion on March 27, 2025. It went public on March 28, 2025, raising $1.5 billion, and was the largest AI-related listing by amount raised, according to Dealogic. The successful listing marked a significant milestone for the company and the broader AI infrastructure sector.
Post-IPO Developments (2025-2026)
In July 2025, CoreWeave became the first company to deploy Nvidia Blackwell Ultra GPUs (GB300 NVL72) commercially, with Dell providing the servers. Core Scientific agreed in July 2025 to be acquired by CoreWeave for $9 billion, but the deal was rejected by Core Scientific shareholders in October 2025. In September 2025, CoreWeave agreed to acquire OpenPipe, a startup specializing in reinforcement learning tools for training AI agents. In October 2025, CoreWeave agreed to acquire Monolith AI, a developer of AI and ML applications for physics.
In January 2026, CoreWeave received $2 billion in investment from Nvidia at a purchase price of $87.20 per share as they expanded their partnership to boost CoreWeave's data center build out. In February 2026, CoreWeave sought $8.5 billion in new financing, using major AI infrastructure contracts with Meta Platforms as collateral.
In April 2026, CoreWeave announced a multi-year agreement with Anthropic to provide Nvidia GPU capacity for production-scale Claude inference workloads, and separately priced an upsized $3.5 billion convertible senior notes offering. These developments came within 48 hours of a separate $21 billion expansion of its infrastructure partnership with Meta. In May 2026, CoreWeave reported first-quarter revenue of $2.08 billion, exceeding analyst expectations of $1.97 billion and more than doubling year-over-year, with revenue backlog reaching nearly $100 billion.
As of June 30, 2026, CoreWeave reported total debt of approximately $35.1 billion in its Q2 2026 quarterly filing, split between recourse (approximately $31.4 billion) and non-recourse (approximately $3.7 billion) facilities structured against hyperscaler contract commitments.
Key People and Employees
Headquartered in New Jersey, as of 2025 the company had around 1,450 employees. Michael Intrator is the company's chief executive and founder, alongside his other cofounders Brian Venturo as chief strategy officer, Brannin McBee as chief development officer, and Peter Salanki as CTO. In 2024, Nitin Agrawal from Google joined as CFO. In August 2024, Sachin Jain, formerly of Oracle's AI department, was hired as chief operating officer. Simultaneously, Chen Goldberg, formerly of Google, was hired as senior VP of engineering. In October 2024, Michelle O'Rourke was named its new chief communications officer.
The leadership team combines experience from traditional finance, technology, and cloud computing sectors. The founders' background as commodities traders informed their early approach to GPU procurement and financing, while later hires from major tech companies brought enterprise-scale operational expertise.