NVIDIA has formed a strategic partnership with IREN to deploy up to 5 gigawatts of AI infrastructure, complemented by a $3.4 billion agreement for IREN to supply AI cloud services for NVIDIA’s internal research. This initiative will prominently feature IREN’s 2-gigawatt Sweetwater campus in Texas, designed to exemplify NVIDIA’s DSX AI factory architecture. As large AI campuses in Texas become increasingly appealing due to space and power availability, the partnership underscores the growing trend of AI factories as essential infrastructure for AI training and service, integrating power, cooling, and compute into a cohesive platform.

IREN: IREN is a data center and power-infrastructure company that has expanded from bitcoin mining into high-density AI compute and cloud services. It is central to the announcement because NVIDIA will use IREN’s Texas campus as a flagship site for a large AI factory deployment and as a supply source for internal AI workloads.
NVIDIA: NVIDIA is a leading semiconductor and AI infrastructure company best known for its GPUs, networking hardware, and enterprise AI software stack. In this news, NVIDIA is partnering with IREN to expand AI infrastructure and to source AI cloud services for its own research workloads, while also gaining a long-term option to invest in IREN equity.
Jensen Huang: Jensen Huang is the co-founder and chief executive officer of NVIDIA, where he has been a prominent spokesperson for the company’s push into AI computing and data center infrastructure. His comments in the news frame AI factories as essential infrastructure and underscore NVIDIA’s strategy around large-scale AI buildouts.

{“AI Factory Trend”: “AI factories are a key infrastructure model for AI systems, integrating power, cooling, networking, and compute into a single comprehensive platform.”, “Texas Deployment”: “Large AI campuses in Texas are favored for hyperscale builds due to the availability of space and power necessary for dense AI infrastructure.”, “Enterprise Demand”: “Enterprise AI agreements increasingly include comprehensive packages of compute capacity alongside software, networking, and operational services.”}