Nvidia Corp. will invest up to $2.1 billion [1] in data-center operator IREN Ltd. to expand artificial intelligence infrastructure.
The deal comes as the tech industry struggles to keep pace with the massive energy and hardware requirements of generative AI. By partnering with a specialized data-center developer, Nvidia aims to accelerate the build-out of next-generation compute power [2].
Announced on May 7, 2026 [4], the agreement includes a five-year option for Nvidia to purchase 30 million IREN shares [3]. This financial arrangement is tied to a broader partnership to deploy up to five GW of AI-focused data-center infrastructure [1].
IREN, which is headquartered in Hong Kong, operates as a global AI data-center developer [5]. The scale of the planned infrastructure is intended to address the soaring demand for AI compute power [2].
Reports on the total value of the partnership vary across sources. While some reports focus on the $2.1 billion investment [1], others value the overall deal at $3.4 billion [5], noting that the $2.1 billion figure represents the share-option component.
Nvidia, based in the U.S., continues to integrate its hardware with large-scale physical infrastructure to maintain its lead in the AI sector. The collaboration with IREN allows the company to secure the physical capacity necessary to house its latest chips.
“Nvidia will invest up to $2.1 billion in IREN.”
This investment signals a shift in Nvidia's strategy from merely providing hardware to securing the physical and electrical infrastructure required to run it. By locking in five GW of capacity and securing share options in IREN, Nvidia is mitigating the risk of 'compute bottlenecks' where chip availability outpaces the availability of power-ready data centers.
