Maximize Your NVIDIA AI Factory Investment with ONES
ONES operates across end-to-end AI factory fabrics and is validated on NVIDIA Spectrum-X, BlueField, InfiniBand, and NVLink technologies. It’s aligned with NVIDIA AI Factory reference architectures and validated with the broader partner ecosystem—storage, servers, host orchestration, and more. Manage Day 0 build/validation, Day 1 tenant operations, and Day 2 telemetry in a single pane.
Skip the browsing. Ask AI.
Hello! How may I help you today?
Press Releases
Ep 20: Networking in the AI Era: SONiC, Open Networking, and AI Ops with Aviz
February 11, 2026
How AVIZ ONES Enables Secure Multi-Tenant North-South Networking in NVIDIA Spectrum-X AI Fabrics
What Experts Are Saying
Gilad Shainer
SVP of networking, NVIDIA
Lightning-fast transformation in groundbreaking technologies such as agentic AI highlight how crucial it is for every business to prioritize networking innovation to gain an edge. The NVIDIA Spectrum-X networking platform enables innovators like Aviz to provide an Ethernet-based multi-tenant AI networking solution that maximizes GPU performance while simplifying operations for enterprises
How It Works
How does ONES make the NVIDIA AI Factory multi-tenant and telemetry-rich?
ONES implements the NVIDIA AI Factory reference architecture end-to-end—across DGX/HGX/NVL systems and the full partner ecosystem (compute, storage, orchestration, NICs, servers, and networking). It brings unified Day 0–2 operational control with multi-tenant awareness and end-to-end telemetry across both networks: the front-end (north–south) for user/app access and the back-end (east–west) for GPU-to-GPU, storage, and distributed training traffic. By validating integrations across the reference design, ONES helps customers run shared AI factories at full potential—improving GPU utilization while reducing operational complexity.
ONES implements the NVIDIA AI Factory reference architecture end-to-end—across DGX/HGX/NVL systems and the full partner ecosystem (compute, storage, orchestration, NICs, servers, and networking). It brings unified Day 0–2 operational control with multi-tenant awareness and end-to-end telemetry across both networks: the front-end (north–south) for user/app access and the back-end (east–west) for GPU-to-GPU, storage, and distributed training traffic. By validating integrations across the reference design, ONES helps customers run shared AI factories at full potential—improving GPU utilization while reducing operational complexity.
Day 0 — Fabric design & validation: Validate the full AI factory (front-end + back-end) against NVIDIA reference architecture before you deploy.
Day 1 — Tenant management: Provision tenants with guardrails—segmentation, quotas, and policy—across the shared AI Factory.
Day 2 — Fabric monitoring & lifecycle: Continuous telemetry, health, upgrades, and change control to keep performance high and operations simple.
Why Use ONES for Your NVIDIA AI Factory?
NVIDIA AI Factory fabrics are powerful — and with ONES, you can operationalize that power across the full factory (front-end + back-end) and at the tenant level for:
1
Day-0 Validation (See it before you deploy):
Validate your AI factory design in NVIDIA AIR or Aviz ONE Center, and generate a proven, custom reference architecture before rollout.
2
Ecosystem-Integrated Operations:
Built to work across the AI factory stack—compute, storage, orchestration, NICs, and networking (including Spectrum-X)—so you don’t stitch tools together.
3
Day-1/2 Multi-Tenant Visibility & Scale:
Tenant isolation and fair usage with end-to-end telemetry (GPU/DPU/platform/network) and lifecycle control to scale without re-architecting ops.
Listen to our latest webinar with NVIDIA to see how ONES makes Spectrum-X truly AI-ready.
See It in Action
Experience ONES in NVIDIA AIR — the official simulation platform for NVIDIA AI Factory. Aligned 100% with NVIDIA’s reference architecture.
Connect with us to schedule a live demo and walk through real multi-tenant AI fabric operations.