Nvidia invested an additional $2 billion in CoreWeave to help the company expand its AI computing capacity. The chip maker purchased CoreWeave Class A common stock at $87.20 per share. CoreWeave plans to add more than 5 gigawatts of AI computing capacity by 2030.
Deal Structure and Terms
According to Bloomberg, the companies announced the investment on Monday. Nvidia already holds a stake in CoreWeave from earlier funding rounds. The partnership extends a relationship that includes a separate service agreement worth more than $6 billion.
Nvidia agreed to buy over $6 billion in services from CoreWeave through 2032. That earlier deal was announced in September 2025. The new investment adds equity ownership to the existing service contract. CoreWeave operates data centers that rent computing power to businesses building AI applications.
Early Access to Products
CoreWeave will be among the first companies to deploy new Nvidia products. The agreement covers storage systems and a new central processing unit. These products have not yet been released to the broader market. Early access gives CoreWeave a competitive advantage in serving AI customers.
Expanding AI Infrastructure
The 5-gigawatt capacity target represents a major expansion of AI computing resources. CoreWeave provides cloud infrastructure specifically designed for AI workloads. The company competes with traditional cloud providers and specialized AI infrastructure firms.
Nvidia dominates the market for AI chips used in data centers. The company makes graphics processing units that power most AI training and inference workloads. Investing in CoreWeave helps Nvidia ensure strong demand for its chips. The partnership also speeds deployment of Nvidia technology at scale.
The investment comes as demand for AI computing capacity continues to grow. Companies building large language models and other AI systems need massive amounts of computing power. CoreWeave focuses on providing that infrastructure with high-performance hardware. The Nvidia investment supports faster buildout of new data centers.
