Mirantis and Netris Unite to Streamline AI Infrastructure with Kubernetes Automation
The demand for scalable and efficient AI infrastructure is surging, but networking complexities often present a significant bottleneck. On Wednesday, March 11, 2026, Mirantis, a Kubernetes-native infrastructure provider for AI, and Netris, a leader in network automation and multi-tenancy for AI, announced a strategic integration designed to address this challenge. This partnership aims to deliver AI infrastructure as a repeatable, multi-tenant cloud product, accelerating time-to-revenue for GPU operators.
The Networking Bottleneck in AI Infrastructure
While GPU availability is no longer the primary constraint, the intricate networking required to support AI workloads remains a major hurdle. Kubernetes, despite its widespread adoption, introduces complexities across services, clusters, infrastructure, and crucially, the networks that connect them. Traditional networking approaches are often manual, fragmented, and slow to adapt, leading to infrastructure debt and hindering scalability. According to Shaun O’Meara, CTO of Mirantis, “In deploying infrastructure for AI, the complexity of the networking is one of the primary challenges.”
A Kubernetes-Native Approach to Networking
The Mirantis and Netris integration fundamentally shifts networking from an afterthought to a core component of Kubernetes cluster delivery. By automating data center networking for both “east-west” (internal) and “north-south” (external) traffic, the integration eliminates manual provisioning and fragmentation. This automation extends to key fabrics like NVIDIA Spectrum-X Ethernet, NVIDIA Quantum-X InfiniBand, and NVIDIA NVLink.
Key Capabilities of the Integration
- Automated Cluster Delivery: Orchestrated delivery of all Kubernetes infrastructure components, including data center networking.
- Predictable AI Performance: Automation of networking for both east-west and north-south traffic to ensure consistent performance.
- Enhanced Tenant Density: Network automation, abstraction, and multi-tenancy with DPU-enabled tenant networking for greater GPU utilization and reduced operating costs.
- Hardware-Enforced Security: Hardware-enforced multi-tenancy with isolation, fault tolerance, and data safety at the switch and DPU level, crucial for regulated and sovereign workloads.
Netris and DPU Integration
Netris is the first commercial networking orchestration platform to integrate NVIDIA BlueField DPUs into the data center network fabric. This integration extends hardware-enforced tenant isolation to the server level, reducing reliance on CPU cores for networking tasks and improving overall efficiency. As Alex Saroyan, CEO and co-founder of Netris, explains, “Netris eliminates that bottleneck by abstracting and automating Ethernet, InfiniBand, NVLink, and BlueField DPUs fabrics.”
Composable Infrastructure and k0rdent AI
The integration reflects Mirantis’ composable infrastructure approach, allowing operators to select networking technologies that meet their specific requirements. Combined with Mirantis’ k0rdent AI, the solution provides a seamlessly integrated, full-stack GPU cloud experience. Mirantis orchestrates the Kubernetes lifecycle, while Netris delivers network automation, abstraction, and multi-tenancy at the hardware layer.
Looking Ahead
This partnership represents a significant step towards simplifying the deployment and management of AI infrastructure. By automating networking and enforcing tenant isolation at the hardware level, Mirantis and Netris are enabling organizations to accelerate their AI initiatives and unlock the full potential of their GPU investments.