NVIDIA Invests Up to US$100 Billion in OpenAI to Build Giant AI Infrastructure

NVIDIA and OpenAI announced a strategic partnership to operate at least 10 gigawatts of NVIDIA systems—a network of AI data centers consisting of millions of GPUs—to support the training and operation of next-generation artificial intelligence models. As part of the agreement, NVIDIA intends to invest up to US$100 billion gradually in OpenAI as each gigawatt begins operation. The first phase is targeted to be active in the second half of 2026 using NVIDIA’s Vera Rubin platform.

Announcement Details

• Infrastructure scale: This partnership is designed to build and operate a minimum of 10 gigawatts of AI data center capacity equipped with NVIDIA systems — a collection of hardware and millions of GPUs intended for very high-level computing needs.

• Investment commitment: NVIDIA intends to disburse up to US$100 billion, released gradually as each gigawatt of infrastructure begins operation. This investment includes support for data center capacity, power resources, and other related operational components.

• Initial schedule: The first gigawatt of NVIDIA systems is expected to start operating in the second half of 2026, using NVIDIA’s platform called Vera Rubin.

Leadership Statements

Jensen Huang, founder and CEO of NVIDIA, highlighted that over a decade of collaboration between the two organizations has driven computing advancements—from DGX supercomputers to breakthroughs like ChatGPT—and regards this investment as the “next big leap.”

Sam Altman, CEO of OpenAI, emphasized that computing is the foundation of the future economy, and the infrastructure resulting from collaboration with NVIDIA will be the foundation for creating new AI breakthroughs that can empower individuals and businesses widely.

Greg Brockman, President of OpenAI, added enthusiasm about the opportunity to operate 10 gigawatts of computing with NVIDIA to push the boundaries of intelligence.

Technical Collaboration and Ecosystem

• OpenAI and NVIDIA will closely collaborate on software roadmap development, model optimization, and hardware-software integration of NVIDIA.

• This partnership complements an existing network of partners—including companies like Microsoft, Oracle, SoftBank, and Stargate—that focus on building world-class AI infrastructure.

• The long-term goal is to support the development of larger models and computing capabilities that bring OpenAI closer to deploying advanced intelligence systems.

Impact and Significance

The proposed investment and infrastructure scale have the potential to:

• Increase the capacity for training and inference of very large AI models, accelerating research and deployment of generative technologies and advanced AI capabilities.
• Shift the economic and technical landscape of AI data centers—expanding the role of hardware vendors (such as NVIDIA) in the AI infrastructure value chain.
• Cause consolidation of computing resources among major players and strategic partners, with implications for access, costs, and regulation related to power usage and large-scale data center construction.

NVIDIA and OpenAI stated they will finalize the details of the initial phases of this partnership in the coming weeks. If realized, the investment commitment of up to US$100 billion and the 10 gigawatt target will be among the most ambitious AI infrastructure developments in industry history—a step that can technically and economically transform the pace of innovation in the field of artificial intelligence.