Aviz Networks and hosted·ai Partner to Simplify Neocloud Deployment and Operations

Partnership combines Aviz AI network automation stack with hosted·ai capabilities, bringing network automation, token visibility, and agentic AI operations to neoclouds

SAN JOSE, Calif.--(BUSINESS WIRE)--Aviz Networks, a software-first, AI networking automation stack, and hosted·ai, the GPU commerce platform for AI infrastructure, today announced a partnership integrating Aviz’s AI factory portfolio, including ONES AI, Aviz Service Gateway, AI Parser, and AI NOC, with the hosted·ai platform. The collaboration combines network automation, token-level observability, and agentic AI operations to help neocloud operators deploy and run AI services.

Every neocloud wants more out of every GPU it runs. hosted·ai is their GPU Commerce Platform: the software to sell GPUs every way the market buys them, and the demand to fill them. Operators sell the same fleet long term, short term, on demand, virtualized or by the token, under their own brand, and hosted·ai brings buyers beyond their own customers. Sharing GPUs across many customers moves the pressure onto the network, and that is where Aviz comes in.

Operating a neocloud requires monitoring, tenant visibility, high-performance network fabrics, and visibility into token consumption and service performance. Many operators still rely on disconnected compute, networking, and monitoring workflows, adding manual work, complicating troubleshooting, and slowing service delivery.

The partnership brings together four complementary capabilities from Aviz’s AI factory portfolio:

ONES AI NVIDIA-certified network automation software that bootstraps, orchestrates, brings multi-tenancy to and operates the neocloud.

Aviz Service Gateway connects isolated neocloud tenants to the internet, clouds, shared services and other data centers. Orchestrated by Aviz ONES, it combines BGP/EVPN routing, VRF/VXLAN termination, NAT, elastic IPs and stateful L3/L4 policy in a software-defined service edge.

AI Parser provides token-level observability into inference usage, latency, and throughput by tenant and application, helping operators understand service performance and consumption.

AI NOC is the agentic AI operations platform that provides customizable SRE skills.

The integration brings Aviz Networks’ networking software layer for NVIDIA AI Factory into the hosted·ai platform, spanning NVIDIA Spectrum X Ethernet, InfiniBand, NVIDIA BlueField, and multi-vendor networking environments, with unified visibility and operations. hosted·ai brings four complementary capabilities to the partnership:

Demand brings buyers to the operator's capacity beyond its own customer base, from enterprises to developers running inference on Packet.ai, and lets operators sell partner capacity beyond their own fleet.

Consumption Models sell capacity long term, short term, on demand, virtualized or by the token, under the operator's own brand.

GPU Hypervisor shares each GPU across up to ten customers, isolated from one another, running unmodified CUDA workloads.

Billing and Payouts meters every GPU hour and token, collects payment and pays out every party automatically.

“Neocloud operators need to turn GPU investments into services customers can rely on. With hosted·ai, we’re bringing Aviz’s AI factory portfolio into the workflows operators use to launch and run their cloud businesses. Our focus is helping them onboard tenants efficiently, understand inference performance, and resolve operational issues as they grow.”
Vishal Shukla, Co-Founder and CEO, Aviz Networks

“Networking is fundamental to the transition from GPU hardware to a fully operational neocloud. Integrating Aviz into the hosted·ai platform lets us bring multi-tenant compute provisioning, networking, isolation, and observability together into a common operating model. That gives our customers a faster path to deploying and monetizing AI infrastructure, from the first GPU cluster through large-scale AI factories.”
Naren Shankar, Co-Founder and Chief Commercial Officer, hosted·ai

The integration extends hosted·ai's platform from compute orchestration into networking, creating a more complete infrastructure foundation for GPU as a service, bare-metal AI infrastructure, Kubernetes environments, inference services, and AI factories. Aviz also brings a multi-vendor networking approach, giving neocloud operators infrastructure choice while maintaining a consistent operating experience across locations and architectures.

About Aviz Networks

Aviz modernizes networking for the AI era with a full-stack, software-first platform that helps neoclouds, sovereign clouds and enterprise accelerate AI factory investments and maximize ROI. Learn more at aviznetworks.com

About hosted·ai

hosted·ai is the GPU Commerce Platform: the software to sell GPUs every way the market buys them, and the demand to fill them. Operators sell capacity under their own brand: long term, short term, on demand or by the token, to their own customers and to buyers from Packet.ai, GPUaaS.com and the hosted·ai GPU Network. Every party is paid out automatically. Founded in 2024, hosted·ai has teams across the US, EMEA and Asia-Pacific. Learn more at https://hosted.ai.