Nvidia invests $2 billion in CoreWeave to accelerate AI data-center buildout to 2030
Nvidia expanded its partnership with CoreWeave with a $2 billion investment, aiming to speed the rollout of “AI factories” and scale computing capacity. The move underscores intensifying competition to secure chips, power and infrastructure as demand for AI services grows.
Nvidia said on Monday, January 26, 2026, that it is investing $2 billion in CoreWeave, deepening a relationship aimed at rapidly expanding specialized AI data centers. The companies said the partnership is designed to accelerate the buildout of more than 5 gigawatts of AI “factories” by 2030 as demand for AI computing continues to surge.
Under the expanded collaboration, CoreWeave will align more closely with Nvidia’s infrastructure roadmap, adopting multiple Nvidia generations and integrating Nvidia platforms across its services. Nvidia said it bought CoreWeave Class A shares at $87.20 per share, signaling confidence in CoreWeave’s cloud strategy and operational plan.
The announcement arrives as the AI industry collides with real-world constraints: power availability, land acquisition, grid interconnections, and the sheer complexity of building and operating high-density compute campuses. For cloud providers, securing reliable energy and infrastructure timelines has become as critical as securing GPUs.
For Nvidia, the deal strengthens a key route to market for its chips by ensuring major capacity is built on Nvidia infrastructure, while also helping lock in an ecosystem of software, networking and platform standards. For CoreWeave, the investment offers a powerful vote of confidence that could help support financing and long-horizon build plans.
The partnership highlights a wider trend in AI: chipmakers, cloud platforms and hyperscalers are tying themselves together with multiyear commitments to ensure supply, capacity and performance. With enterprise and consumer AI workloads still rising, the competition is shifting from “who has the best model” to “who can reliably deliver compute at scale.”