At IREN, we believe AI builders shouldn't have to choose between raw performance and operational flexibility. As an inaugural partner of the NVIDIA AI Cloud Ready Initiative, Mirantis has completed NVIDIA-Certified Hypervisor validation for k0rdent AI’s virtualization stack on Arm-based NVIDIA GB200 systems. This milestone demonstrates that virtualized GPU environments can now deliver predictable, near bare-metal performance while supporting secure, hardware-level multi-tenant isolation. IREN AI Cloud empowers engineering teams to match infrastructure to workload requirements, balancing performance, isolation, and operational efficiency. AI builders can choose from: • Software-Defined Virtualization: delivers secure, elastic multi-tenant environments and scalable inference, with the VM as a hardened tenancy boundary. • Dedicated Bare-Metal Clusters: optimized for maximum performance to deliver high-throughput model training and high-concurrency, ultra-low-latency inference. From infrastructure to inference, choose the operational model that best fits every AI workload.