For companies ready to build a large-scale AI Factory, NVIDIA DSX provides a platform that unifies design, simulation, operations, and ecosystem technologies that optimize tokens per watt. Embracing this technology, ASUS is ready to help customers integrate NVIDIA DSX-ready systems into real-world deployment plans — from early facility assessment to infrastructure buildout, operations, and scalable inference — with the goal of helping customers reduce deployment risks and accelerate the path from AI factory planning to real-world buildout.

Accelerating AI Factory Buildout through Power and Facility Planning

For customers building large-scale AI factories, one of the first questions is simple but critical: how much power will the data center need, and can the facility support it?

This is not only about total power capacity. Customers also need to evaluate power distribution, rack-level power delivery, uninterrupted power supply (UPS), and backup power, generator readiness, networking configuration, floor loading, service clearance, and future expansion requirements. Any oversight in these areas could lead to costly redesign, delayed installation or additional facility work, leading to higher buildout costs.

The NVIDIA Omniverse DSX Blueprint and NVIDIA DSX SimReady are crucial resources that help with all aspects of the planning phase. The NVIDIA Omniverse DSX Blueprint provides the digital twin framework where AI factory layouts, facility conditions, and infrastructure behavior can be designed, simulated, and validated before physical buildout; NVIDIA DSX SimReady enables facilities to be designed, verified, and optimized in simulation before physical buildout. By converting rack-scale systems into validated OpenUSD assets with embedded power, thermal, and connection-point metadata, ASUS and relevant hardware suppliers can be represented inside NVIDIA Omniverse DSX Blueprint as simulation-ready assets.

Leveraging accurate simulations from DSX ecosystem tools, customers can better understand whether the planned AI factory can support the expected compute capacity, power demand, cooling behavior, and connection requirements. ASUS can then help translate these insights into a detailed plan using the ASUS Infrastructure Deployment Center (AIDC) — a proprietary automation engine for rapidly deploying AI clusters at scale.

Optimizing Cooling Design, Water Conditions, and Energy Cost

Power planning is only one part of AI factory readiness. For high-density systems, cooling design directly affects reliability, energy efficiency, and operating cost.

By developing with the NVIDIA Omniverse DSX Blueprint, ASUS can accurately design cooling in the early stages of AI factory planning. By using NVIDIA DSX SimReady to connect ASUS rack-scale infrastructure with ecosystem cooling partner models, the digital twin can

reflect compute systems as well as cooling conditions and facility constraints that affect real-world deployment.

This allows ASUS customers to evaluate cooling assumptions as part of the overall AI factory plan, rather than treating cooling as a separate facility issue. For example, teams can better understand how water temperature, water flow, water pressure, Cooling Distribution Units (CDUs) placement, piping design, and rack-level heat load may influence system stability, energy consumption, and long-term cost before equipment arrives on site. Vera Rubin 3rd generation MGX racks are engineered for 45°C warm-water inlet temperatures, enabling ambient-air dry-cooler designs that drive down PUE and redirect power from cooling overhead into token generation.

Through this approach, ASUS can help customers turn NVIDIA AI factory blueprint into a more concrete AI factory deployment plan. By aligning rack-scale system design, liquid-cooling architecture, cooling partner inputs, and facility conditions, ASUS helps reduce the risk of under-provisioning or costly reworking of systems.

Reliable Long-Term System Monitoring

After an AI factory is deployed, the challenge shifts from buildout to long-term operations, which involves accurate monitoring. Proprietary ASUS Control Center Data Center Edition (ACC) software gives customers continuous visibility across system health, power status, cooling conditions, networking behavior, workload changes, and maintenance impact to keep large-scale AI infrastructure running reliably. If monitoring indicates that adjustments need to be made — such as workload balance, cooling, rack configuration, or infrastructure expansion — customers can use NVIDIA DSX to simulate and validate revisions before applying them in the physical environment.

This is where NVIDIA DSX Exchange comes in. As an AI Factory IT/OT communication hub, it links and tracks compute, networking, energy, power, and cooling data across all relevant systems. Using insights from the data collected via Exchange, customers can better understand how infrastructure changes may affect system stability, service continuity, and long-term efficiency.

Once the optimal adjustment plan is confirmed, ASUS AIDC can help streamline deployment workflows and accelerate implementation.

Scaling Inference Capacity with Better Performance Per Watt

As AI applications move from pilot projects to production, customers may need to scale up inference capacity quickly.

However, scaling inference is not simply about adding more systems. Customers also need to understand whether the existing power budget, cooling capacity, and infrastructure design can support the expansion without increasing cost or disrupting current services.

With NVIDIA DSX MaxLPS, customers can evaluate how to maximize compute output and token performance per watt on NVIDIA systems within a fixed power budget. NVIDIA Dynamo further supports scalable inference serving across distributed AI infrastructure, helping production AI services handle growing request volumes efficiently.

ASUS AIDC is ready to help customers turn their scaling strategy into a practical infrastructure plan. By aligning power-aware planning, liquid-cooling design, networking and storage integration, deployment workflows, and management capabilities, ASUS helps facilitate every aspect of AI factory capacity expansion in a streamlined, efficient way.

Building the Next Step of AI Factory Readiness

By embracing NVIDIA DSX’s AI factory-scale vision, ASUS helps customers turn digital twin insights into deployment-ready infrastructure plans for AI factories built with systems such as ASUS AI POD with NVIDIA Vera Rubin NVL72.

After deployment, ASUS continues to provide add-on value through ASUS Infrastructure Deployment Center, ASUS Control Center, liquid-cooling expertise, professional services, and ecosystem collaboration.

Explore how ASUS can help you turn NVIDIA’s AI factory blueprint into real-world AI factory