IREN Secures $2.8B in AI Cloud Contracts, Raises 2026 Revenue Target Above $4B
Key Takeaways
- •IREN has closed three AI cloud service contracts with a combined value of $2.8 billion, covering bare-metal GPU infrastructure and managed AI cloud services.
- •The company raised its annualized run-rate revenue target to over $4 billion, with approximately 85% of that figure already secured under existing contracts.
- •IREN plans to expand its AI cloud capacity from 480 MW targeted for this year to 1.2 GW by 2027, a significant infrastructure buildout requiring substantial GPU and electrical resources.
- •The latest contract includes a prepayment covering roughly 45% of GPU deployment costs, reducing IREN's capital requirements and accelerating its growth trajectory.
- •IREN's strategic pivot mirrors a broader industry trend in which cryptocurrency mining firms are reallocating power-intensive infrastructure toward AI workloads for more predictable revenue streams.

IREN, a company originally established as a Bitcoin mining firm, has closed three new AI cloud service contracts with a combined value of $2.8 billion, significantly deepening its position in the artificial intelligence infrastructure market. The deals, signed with leading AI developers, cover both bare-metal GPU infrastructure and managed AI cloud services.
Revised Revenue Outlook
Following the multi-year agreements, IREN now projects its annualized run-rate revenue (ARR) to surpass $4 billion, up from a previous forecast of $3.7 billion and well above the current ARR of more than $2.5 billion. The company stated that approximately 85% of the new ARR target is already secured under contract, offering considerable revenue visibility as additional AI infrastructure is deployed.
IREN's customer portfolio now includes Microsoft, NVIDIA, Perplexity, Figure AI, Together AI, Fluidstack, Fireworks AI, Fal AI, Hume AI, and one additional leading AI developer.
According to an SEC filing, the company has formalized these multi-year cloud agreements to meet growing demand from hyperscalers, enterprises, frontier AI labs, and AI developers.
Capacity Expansion and Deal Terms
IREN reports that demand for AI computing capacity continues to outpace its available portfolio. The company's AI cloud capacity has scaled rapidly — from roughly 3 MW one year ago to a target of 480 MW this year — with plans to reach 1.2 GW by 2027. Scaling from 480 MW to 1.2 GW in roughly two years represents a significant infrastructure buildout that will require substantial GPU procurement, electrical capacity, and cooling infrastructure.
The latest contract includes a prepayment covering approximately 45% of GPU deployment costs, which IREN says reduces its capital requirements and accelerates growth. The average duration of the company's AI contracts is approximately four years.
Strategic Transition to AI Infrastructure
Founded as a Bitcoin mining business, IREN is steadily repositioning itself as a diversified AI infrastructure provider. Its data centers, located in renewable-energy-rich regions across North America, Europe, and APAC, are designed to support AI model training and inference as well as digital asset operations. Access to renewable energy is increasingly relevant as large AI customers face growing pressure to report and reduce the carbon footprint of their compute-intensive workloads.
As of June 30, 2026, IREN reported a cash and cash equivalents balance of approximately $7.6 billion, which includes restricted cash tied to GPU-related financing arrangements.
Broader Industry Trend
IREN's pivot reflects a wider movement across the cryptocurrency mining sector. As mining margins have grown more competitive, companies are leveraging their existing power infrastructure and data center capabilities to serve surging demand for AI compute. Firms such as Core Scientific and Hut 8 have pursued similar transitions, reflecting a broader reallocation of power-intensive infrastructure toward AI workloads. Long-term cloud contracts offer these firms more predictable revenue streams compared to Bitcoin mining, while allowing enterprise customers to benefit from their large-scale computing facilities.