AI Demand Extends Into 2027
IREN said it is being selective about how it assigns capacity before new infrastructure comes online. The company is seeking a mix of customers across hyperscalers, enterprises, AI developers and advanced research labs.
Demand remains higher than IREN’s existing and planned capacity. The company is already in discussions with customers for infrastructure scheduled to become available throughout 2026 and 2027.
Over the past year, IREN has expanded from about 3 megawatts of self-built AI Cloud capacity. It now expects to deliver 480 MW during 2026 and is targeting 1.2 gigawatts in 2027.
Co-founder and co-CEO Daniel Roberts said the platform was scaling quickly as IREN broadened its customer base. He commented:
We are proud to support leading companies building frontier applications across design, physical AI and robotics, generative media, AI search and model development.
Prepayments Lower Capital Requirements
IREN also said pricing on new contracts continues to improve. Some customers have agreed to make upfront payments equal to about 45% of the capital cost of the graphics processing units tied to their deployments.
Those payments reduce the amount of capital IREN must provide before the services begin generating revenue.
Across its portfolio, the company’s customer contracts have a weighted average term of about four years. That gives IREN greater revenue visibility as it expands its data center network. As of June 30, 2026, the company held about $7.6 billion in cash and cash equivalents.