Market Reaction: CoreWeave Shares Strengthen as Investors Re-evaluate AI Cloud Demand
Boosted by news of Nvidia's increased investment, AI cloud service provider CoreWeave surged by about 10% during intraday trading on Monday, seen by the market as a signal that the "compute power expansion cycle is still accelerating."
Investment Details: $2 Billion Stake, Priced at $87.20 Per Share
According to disclosures from both parties, Nvidia will purchase CoreWeave's Class A common stock at $87.20 per share, with a total investment amounting to approximately $2 billion. Reuters also confirmed the investment amount and price, noting that the partnership upgrade aims to support CoreWeave's data center expansion plans.
Collaboration Focus: Advancing Over 5GW "AI Factory" Construction by 2030
The main mission of this collaboration is to rapidly and substantially increase computing power supply: the two companies aim to facilitate the construction of AI factories exceeding a 5-gigawatt scale (understood as ultra-large-scale data center clusters for AI training and inference) by 2030, with CoreWeave responsible for development and operation, while Nvidia provides accelerated computing platforms and ecosystem support.
On the resources front, Nvidia will also leverage its capital and industry synergies to help CoreWeave quickly acquire critical elements like land, electricity, and "skeletons" for data centers needed for plant construction, thereby shortening the production cycle.
Technology and Products: Binding Roadmaps Further from Software to Next-Gen Platforms
Besides “building faster,” both parties also emphasize ease of use. The partnership includes testing and validating CoreWeave's AI-native software and reference architecture (such as SUNK and CoreWeave Mission Control) with the goal of enhancing interoperability and aiming for inclusion in Nvidia's reference architecture system for cloud partners and corporate clients.
In terms of hardware roadmaps, CoreWeave plans to pre-integrate multiple generations of Nvidia's computing architecture into its platform, including the Rubin platform, Vera CPU, and BlueField storage systems, to maintain scaling flexibility as training and inference demands change rapidly.
