Tech giant NVIDIA has announced a USD 2 billion equity investment in AI cloud-computing company CoreWeave, highlighting the accelerating pace of infrastructure deals in the AI sector.

Under the agreement, NVIDIA will acquire Class A common stock in CoreWeave and deepen a technical and commercial partnership aimed at speeding up the construction of specialised data centres, often referred to as “AI factories”. CoreWeave has set a goal of developing more than 5 gigawatts of capacity by 2030 to support growing demand for AI workloads.

“AI is entering its next frontier and driving the largest infrastructure buildout in human history,” said Jensen Huang, Founder and CEO of NVIDIA. He added that CoreWeave’s experience in building AI-focused data centres and its execution capabilities were well recognised across the industry, as both companies work to address rising demand for AI infrastructure.

The arrangement combines direct capital investment, closer product integration, and shared exposure to future demand for hardware and cloud services. This structure reflects the increasing strategic importance of access to large-scale computing capacity as AI models become more complex and resource-intensive.

As part of the collaboration, NVIDIA and CoreWeave will increase coordination across both infrastructure and software. NVIDIA will assist CoreWeave in securing land, power supply, and data centre facilities, while CoreWeave’s software platform will be tested and validated to ensure closer interoperability. There is also the possibility that CoreWeave’s software could be incorporated into NVIDIA’s reference architectures.

CoreWeave plans to deploy multiple generations of NVIDIA hardware across its infrastructure stack. The partnership also aims to expand the availability of CoreWeave’s software to global cloud service providers and enterprise customers, extending its reach beyond its existing user base.

The transaction reflects a broader industry trend often described as a “capital-plus-compute” partnership. In such deals, technology vendors invest capital or take equity stakes in cloud operators or AI developers, while securing long-term product integration and, in some cases, guaranteed demand for computing capacity.

Several recent agreements follow a similar pattern. Microsoft and NVIDIA have committed significant investment tranches to Anthropic, which in turn agreed to multi-year cloud purchases. Around the same time, Anthropic expanded its use of Google Cloud’s TPU infrastructure, illustrating how AI developers often rely on multiple partnerships to secure capacity and pricing.

Other companies are taking alternative approaches. Meta has prioritised large-scale internal infrastructure development, committing to extensive GPU inventories and multi-gigawatt data centre projects to retain control over compute resources. Earlier this month, ChatGPT creator OpenAI announced a multi-year partnership worth over USD 10 billion with AI chipmaker Cerebras to strengthen its computing infrastructure.