NVIDIA and IREN: 5 GW to Revolutionize Global AI Infrastructure

NVIDIA and IREN Join Forces for AI

NVIDIA, the global leader in computing technologies, announced on May 7 an ambitious partnership with IREN, aiming to deploy up to 5 gigawatts of infrastructure dedicated to artificial intelligence (AI). This project is based on NVIDIA's DSX architecture and will extend across IREN's global network of data centers. The goal is to meet the exponential increase in computing power demands from businesses and AI specialists.

A Global Scale Project

This strategic partnership between NVIDIA and IREN marks a significant milestone in the global race for computing power. By combining NVIDIA's GPUs, energy, and IREN's data centers, the two companies aim to accelerate the deployment of AI factories capable of meeting the growing demand from businesses and AI players. The Sweetwater campus, located in Texas, is designated as the flagship site for this deployment. With a capacity of 2 gigawatts, this site benefits from privileged access to energy, available land, and robust industrial infrastructure. These advantages make Texas a prime destination for investments in AI and data centers.

The Roles of NVIDIA and IREN

In this partnership, NVIDIA brings its expertise in accelerated computing and AI architecture, while IREN manages critical infrastructure aspects, such as energy, land, and data center operations. The project primarily targets AI-native companies, specialized startups, and large groups requiring significant computing capabilities for training and inference of models.

A Strategic Repositioning for NVIDIA

Historically, competition in the AI field focused on model performance. Today, the competitive advantage also lies in the ability to rapidly build infrastructures capable of supporting these systems. Jensen Huang, founder and CEO of NVIDIA, now views AI factories as fundamental infrastructures, comparable to electrical networks or telecommunications. This vision confirms NVIDIA's gradual repositioning, which no longer just sells GPUs but now offers a complete infrastructure for industrial-scale AI.

The Texas Sweetwater Campus at the Heart of the Project

Future deployments are expected to focus primarily on IREN's Sweetwater campus in Texas. This 2-gigawatt site is the future flagship project for NVIDIA's DSX architecture. The U.S. state is increasingly attracting investments in AI and data centers due to its access to energy, available land, and industrial infrastructure. For companies in the sector, energy considerations are becoming as important as access to semiconductors.

IREN has a unique positioning in this market. The company is developing a vertically integrated model that combines data centers, access to electrical networks, and operation of GPU clusters intended for AI training and inference. Its platform relies on infrastructures located in regions rich in renewable energy across North America, Europe, and Asia-Pacific. This aspect is becoming strategic as the energy needs of AI models increase.

The Financial Implications of the Agreement

The agreement between NVIDIA and IREN includes a purchase right for NVIDIA on up to 30 million shares of IREN's common stock, at an exercise price of $70 per share. This potential investment could reach $2.1 billion, thereby strengthening NVIDIA's position in the AI value chain.

A Strategic Alliance Confirming the Industrialization of AI

After the race for generative models, the industry is now entering a phase of massive industrialization of infrastructures. The demand for computing power continues to rise with the proliferation of AI agents, multimodal models, and real-time applications. Hyperscalers, cloud providers, and specialized players are therefore seeking to secure energy capacities and data centers that can support this growth over several years.

Thus, for NVIDIA, this alliance with IREN allows it to strengthen its control over the entire AI value chain, including hardware, software, networking, and now energy and operational infrastructure. For IREN, the agreement brings additional credibility in an increasingly competitive market. The company is now positioning itself as an infrastructure partner capable of meeting the needs of AI giants. According to the announcement made by NVIDIA, both groups aim to accelerate global access to AI computing power and reduce the deployment times of critical infrastructures, a central issue as companies multiply large-scale AI projects.