Network Automation for Bare Metal Management: our Partnership with Aviz Networks
CLASTIX and Aviz Networks are partnering to bring network automation in Bare Metal management, helping GPUs cloud providers, neoclouds and enterprises deploy and run multi-tenant bare metal infrastructures with less manual work. CEO Adriano Pezzuto on why compute and network need to work as one, and what it means for operators turning bare metal into revenue.

- Written by

FOUNDER · CEO
- Published on
- November 10, 2026
- Read Time
- 3 min
Why bare metal management is hard at scale
Bare metal management sounds simple: just give me the hardware and access the raw capacity. But running it at scale is a completely different engineering problem. Preparing one bare metal server manually is manageable. A senior system engineer can quickly install the hardware components, provision the operating system, install the required drivers, and prepare the machine for use.
At fleet scale, those steps have to be executed consistently across hundreds of physical servers, network devices and multiple tenants. Server state, software versions, network configuration, and tenant isolation must remain consistent as machines and network devices are provisioned and reprovisioned several times.
In other terms, Bare Metal Management is turning raw capacity into multi-tenant, production-ready environments that customers can actually use, and doing it quickly enough to keep pace with demand.
Compute and network: two layers that have to work as one
Getting there means coordinating two layers that have traditionally lived apart. On one side is compute: provisioning Kubernetes environments and allocating bare metal machines for tenants. On the other is the network: establishing secure and isolated connectivity for each tenant across the entire network fabric. In most organizations these are handled by different teams, with different tools, through a series of manual steps. Every one of those steps adds time between buying new hardware and earning revenue from it.
CLASTIX and Aviz Networks: network automation for AI factories
That gap is exactly what we set out to close. Today I’m glad to share that CLASTIX is partnering with Aviz Networks to develop network automation for AI factories. Read the full announcement here: Aviz Networks and Clastix Partner to Automate AI Networking.
Together we’re bringing our Kubernetes platform capabilities together with the Aviz AI Factory stack, Aviz’s NVIDIA-certified network automation software, to help operators deploy and run multi-tenant infrastructure with far less friction.
How kMetal and the Aviz AI Factory stack fit together
The two sides fit together naturally. Through kMetal, our platform for provisioning and managing Kubernetes on bare metal, we remove much of the per-cluster overhead that comes with operating Kubernetes at scale. Aviz automates the networking that connects those environments. As an NVIDIA Inception program member working with an NVIDIA-certified partner, we share a common foundation in the NVIDIA ecosystem, which makes this integration all the more natural for the GPUs cloud providers, neoclouds and enterprises we both serve.
What this means for multi-tenant GPU infrastructure
For the operators building this infrastructure, the partnership is designed to deliver three things:
- Faster time to online. Tighter coordination between compute setup and network provisioning, so bare metal environments come up sooner.
- Multi-tenant connectivity. Clear, separate network connectivity for each tenant environment.
- Less manual work as you grow. Automated network configuration and operations that scale with the infrastructure rather than against it.
A cloud-like experience on bare metal you control
This reflects what we keep hearing that our customers want: the flexibility of running bare metal infrastructure they control, with the simplicity of a cloud-like experience. Bringing network automation closer to Kubernetes operations is a meaningful step toward that, and toward helping operators turn bare metal capacity into revenue sooner.
We’ll share more on availability and supported configurations as the collaboration progresses. In the meantime, if you’re building multi-tenant bare metal infrastructure for AI factories and want to move faster, talk to us.