Aviz Networks partners with Spectro Cloud to unify operations for NVIDIA AI infrastructure
Aviz Networks and Spectro Cloud established a collaboration aimed at providing a unified operational framework for NVIDIA Artificial Intelligence (AI) infrastructure. The alliance focuses on integrating GPU-aware Ethernet fabrics with Kubernetes fleet lifecycle automation and alignment with NVIDIA AI Enterprise software. This initiative supports consistent performance management from initial deployment through subsequent operational phases.
The partnership targets the operational complexities encountered when scaling Graphics Processing Unit (GPU) clusters, emphasizing the necessity for integrated management spanning networking, compute, and storage elements. The cooperative effort addresses the alignment of infrastructure topology with cluster lifecycle processes and governance requirements within Kubernetes and NVIDIA hardware environments.
Aviz Networks contributes deterministic fabric orchestration capabilities that include GPU-aware topology alignment and multi-tenant segmentation across Ethernet fabrics. Spectro Cloud brings its PaletteAI platform, which manages full-stack deployment and operations of AI environments incorporating NVIDIA AI Enterprise components. Together, these components provide automation and orchestration functions designed for AI Factory Infrastructure as referenced in the NVIDIA architecture.
The combined solution standardizes Kubernetes environments with verified AI stack blueprints, optimizes Ethernet fabric performance to enhance GPU cluster efficiency, enforces multi-tenancy controls with policy guardrails, automates lifecycle operations from initial deployment through ongoing management, and integrates observability across networking and compute layers for improved troubleshooting and resource utilization.
Leon Jones, vice president of alliances at Spectro Cloud, said, “Cloud providers and enterprises want to consume GPU infrastructure like a product without reinventing orchestration and operations for every cluster. With PaletteAI and Aviz ONES together, customers can standardize AI stack deployment, automate lifecycle management, and operate AI factories with consistent performance and governance across environments.” Thomas Scheibe, chief product officer at Aviz, said, “Aviz was founded to make AI networking simple, open, and multi-vendor while giving infrastructure teams a deterministic way to operate AI fabrics at scale. With Spectro Cloud, we're delivering a unified operations stack that aligns Kubernetes fleets, GPU-aware Ethernet fabrics, and multi-tenant segmentation so AI factories can be deployed faster, run predictably, and be consumed securely like true infrastructure-as-a-service.”
The collaboration is positioned to support AI platform operations that require governed, multi-tenant AI clusters with policy-driven controls; deployment of AI Factory environments with deterministic GPU infrastructure; GPU Platform-as-a-Service (PaaS) environments delivering isolated AI instances; neo-cloud and GPU providers offering infrastructure catalog services; and hybrid and sovereign AI operators integrating operations across regulated contexts.