IREN Signs $2.8 Billion in AI Cloud Contracts and Lifts 2026 ARR Target Above $4 Billion

IREN said it has signed $2.8 billion in new multi-year AI cloud services contracts with leading AI developers and raised its 2026 AI Cloud annualized recurring revenue target to more than $4 billion.

IREN raises its AI cloud revenue outlook

In a July 20 GlobeNewswire release, IREN said the newly signed customer contracts represent $2.8 billion in total contract value. The company said the agreements are for AI cloud services and involve “leading AI developers,” though the release excerpt did not identify the customers by name.

IREN also said the new contracts lift its 2026 AI Cloud ARR target to over $4 billion. ARR is commonly used by subscription and services businesses to describe annualized recurring revenue based on contracted or recurring customer commitments.

The announcement positions IREN among infrastructure companies seeking to convert demand for graphics processing unit capacity into longer-term cloud services revenue. The company’s release, published by GlobeNewswire, frames the contracts as part of its AI cloud business rather than its historical association with energy-intensive computing.

AI infrastructure demand is broadening beyond raw GPU supply

IREN’s update arrives alongside other infrastructure announcements that show AI cloud demand is increasingly tied to networking, deployment geography, and production monitoring.

CoreWeave said in a company blog post that AI cloud networking has become strategic for GPU “goodput” in training and inference. The company said liquid-cooled switches can help remove networking bottlenecks in large-scale AI systems and can double AI network bandwidth per rack. That focus suggests cloud providers are competing not only on the number of accelerators available, but also on how efficiently those accelerators can be used.

Nebius separately announced an expansion in the United Kingdom, describing about £1.7 billion of AI infrastructure investment tied to new NVIDIA-powered deployments. In its announcement, Nebius also pointed to production inference use cases, including Revolut’s AI stack and Nebius Token Factory. The company’s update indicates that demand is coming from enterprise and agentic AI applications as well as model development.

Amazon Web Services has also emphasized the operational side of production AI. AWS announced a new Amazon SageMaker AI inference observability capability for inference endpoints, saying it exposes metrics such as token latency, GPU health, placement, autoscaling, queue depth, and throughput. Those metrics are relevant for companies running generative AI services where latency, reliability, and cost control can determine whether an application is usable at scale.

What to watch

IREN’s $2.8 billion contract announcement is significant because it gives the company a larger stated revenue target for its AI cloud business. The key questions now are execution-related: how quickly IREN can bring contracted capacity online, whether customers expand or renew commitments, and how margins evolve as AI infrastructure providers invest in power, cooling, networking, and hardware.

The broader market context also matters. CoreWeave’s focus on liquid-cooled networking, Nebius’s regional expansion, and AWS’s inference observability tools all point to a maturing AI infrastructure market where customers are looking beyond accelerator access alone. For IREN, the new contracts raise the stakes for delivery in 2026.