Aviz Integrates with NVIDIA DSX Air to Cut Validation Time
At the NVIDIA GTC conference on March 16, 2026, Aviz Networks announced the expanded availability of its ONES validation platform within NVIDIA DSX Air. The integration allows customers and partners to create cloud-based digital twins of NVIDIA AI Factory networking designs. This enables them to simulate and validate the entire complex infrastructure stack—from networking and DPUs to storage and software—before committing to physical deployment, significantly reducing risk and accelerating time-to-operation.
This 'shift-left' approach to validation is critical for the fast-evolving world of AI infrastructure. By testing designs in a virtual environment, enterprises can identify and resolve potential bottlenecks and integration issues early, ensuring that the physical AI factory performs optimally from day one. This capability is crucial for companies like CoreWeave, which already uses NVIDIA DSX Air to test digital twins of its AI factories and shorten validation times.
Integration Supports NVIDIA's New AI Factory Blueprint
The Aviz partnership is a key component of NVIDIA's broader strategy to standardize the construction of large-scale AI infrastructure. On the same day, NVIDIA unveiled its Vera Rubin DSX AI Factory reference design and the generally available Omniverse DSX Blueprint. These initiatives provide a detailed guide for designing, building, and operating the entire AI factory stack for repeatable and scalable performance. The goal is to create an open, modular software stack that maximizes the output of AI tokens per watt of energy consumed.
In the age of AI, intelligence tokens are the new currency, and AI factories are the infrastructure that generates them.
— Jensen Huang, Founder and CEO of NVIDIA.
By providing a framework that unifies power, cooling, networking, and compute, NVIDIA and its ecosystem partners, including Aviz, are working to transform the complex process of building an AI data center into a more streamlined, product-level system. This effort has attracted broad industry support from leaders like Siemens, Schneider Electric, and Dell.
Tackling the $300B AI Infrastructure Challenge
The push for standardized designs and faster validation directly addresses the immense challenges facing the AI industry. Energy has become the primary bottleneck for AI infrastructure buildouts, with industry reports citing over $300 billion in backlogged equipment orders and more than 200 gigawatts of projects waiting in U.S. interconnection queues. The complexity of coordinating power, cooling, and compute for these massive facilities makes traditional design methods slow and inefficient.
Platforms like NVIDIA DSX Air, now enhanced with Aviz's networking tools, provide a solution by enabling physically accurate simulation. This allows builders and operators to optimize layouts, power topologies, and thermal behavior before breaking ground. By de-risking the construction phase and ensuring maximum efficiency, NVIDIA and its partners aim to unlock faster access to the computational power needed to meet the surging global demand for AI.
