Nvidia pours $2B into CoreWeave to build 5 GW of AI factories
NVIDIA expands partnership with CoreWeave to accelerate buildout of over 5 GW of AI factories by 2030, leveraging its financial strength and platform technology. NVIDIA invested $2 billion in CoreWeave Class A common stock at $87.20 per share, underpinning confidence in CoreWeave’s cloud AI growth strategy.
1. Expanded Strategic Partnership with CoreWeave
NVIDIA has committed a $2 billion equity investment in CoreWeave Class A common stock at $87.20 per share to deepen a multi-year infrastructure alliance. This infusion bolsters CoreWeave’s balance sheet and underlines NVIDIA’s conviction in CoreWeave’s cloud-platform model. The capital will be deployed to accelerate site acquisition—land, power and shell—and to secure long-lead equipment orders for more than 5 gigawatts of AI-optimized data centers by 2030, supporting hyperscale AI workloads for global enterprise and cloud customers.
2. Joint Buildout of AI Factories at Scale
Under the expanded collaboration, NVIDIA and CoreWeave will jointly develop and operate purpose-built “AI factories” powered by NVIDIA’s Rubin accelerated compute architecture, upcoming Vera CPUs and BlueField storage accelerators. CoreWeave will manage facility design, construction and operations, while NVIDIA will contribute early access to next-generation GPU, CPU and networking platforms. Both companies will co-validate CoreWeave’s AI-native software stack, including SUNK orchestration and Mission Control reference architectures, to ensure turnkey integration for enterprise AI deployments.
3. Technology Alignment and Reference Architecture Integration
Together, NVIDIA and CoreWeave will align hardware and software roadmaps to deliver multi-generation interoperability. CoreWeave will test NVIDIA’s Rubin GPU clusters, Vera central processors and BlueField Data Processing Units in production environments, feeding performance data back into NVIDIA’s reference architecture libraries. This initiative paves the way for inclusion of CoreWeave-validated infrastructure blueprints in NVIDIA’s partner ecosystem, giving enterprise customers and cloud providers standardized, interoperable designs for AI training, inference and data preprocessing workloads.
4. Implications for Compute Demand and Market Position
This collaboration positions NVIDIA to capture an even larger share of the rapid compute-infrastructure expansion driven by generative AI and large-scale machine-learning models. By jointly deploying over 5 gigawatts of AI-optimized capacity, the companies anticipate meeting surging demand from hyperscalers, research labs and enterprise clients over the next decade. The partnership leverages NVIDIA’s balance-sheet strength and CoreWeave’s operational expertise, strengthening NVIDIA’s leadership in the AI infrastructure market and setting a new benchmark for end-to-end data-center solutions.