ASUS Ramps Up NVIDIA Vera Rubin NVL72 Shipments as Demand for AI Infrastructure Grows

ASUS is increasing volume shipments of its NVIDIA Vera Rubin NVL72 systems in the fourth quarter of 2026, responding to rising demand for large-scale artificial intelligence infrastructure. The company is also showcasing an integrated AI Factory reference architecture at SC26 in Chicago, highlighting how it plans to support customers with rack-scale computing, automated manufacturing, infrastructure planning and deployment services.

The announcement signals a broader shift in the AI server market, where the ability to supply powerful computing systems is becoming only one part of the equation. Enterprises and cloud providers building large AI environments also need reliable rack-level integration, advanced cooling, high-speed networking, storage and support throughout deployment and operations.

ASUS says its expanded production capacity is backed by automated L10 and L11 manufacturing processes, engineering-led risk assessment and a six-layer validation framework. These measures are intended to make deliveries more consistent and reduce the technical risks involved in installing complex AI systems.

The company is presenting its solutions at Booth 2234 at SC26, held at McCormick Place in Chicago.

ASUS Vera Rubin NVL72 Anchors Its AI Factory Architecture

At the centre of the showcase is the ASUS AI POD XA VR721-E3, a rack-scale system built around NVIDIA’s Vera Rubin platform. The architecture integrates 72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs in a fully liquid-cooled configuration.

The system is designed for demanding AI workloads, including frontier-model training, complex reasoning and high-throughput inference. Its configuration also incorporates NVIDIA Groq 3 LPX, Vera CPU trays, the NVIDIA Spectrum-X SN6600 Ethernet switch and NVIDIA CMX context memory technology.

By bringing computing, networking and memory components together within a rack-scale architecture, ASUS is positioning the AI POD as a foundation for organisations building large AI computing environments.

The company is also displaying two NVIDIA HGX Rubin NVL8-based systems. These include the 3U hybrid-cooled XR R1I12B-M1BHFS and the 2U fully liquid-cooled XR R1I12B-M1BLFS, offering customers different cooling and deployment configurations.

Its wider NVIDIA MGX portfolio includes the XA P2N-E2, powered by NVIDIA Vera CPUs and designed for multi-step agent execution and orchestration. Another system, the ESC8000A-E13P, uses NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs for professional accelerated computing.

Automated Manufacturing and Validation Support Rack-Scale Deployments

Building large AI systems involves more than assembling individual servers. Rack-scale deployments require precise coordination across power delivery, thermal management, high-speed interconnects and system integration.

ASUS says it uses its Predictive Virtual Lab to identify potential engineering risks, including signal-integrity issues, during the design stage. Identifying these problems early can help reduce complications when systems move into manufacturing and deployment.

The company is also expanding automation across its L10 and L11 manufacturing processes. These include automated optical inspection, protective-cover removal, module-positioning assistance and precision fastening. The aim is to reduce assembly variations and improve consistency as production volumes increase.

ASUS has introduced a six-layer validation framework covering both individual nodes and complete racks. The process includes hardware functionality checks, BIOS and baseboard management controller verification, power cycling, stress testing, NVIDIA diagnostics and rack-level performance evaluation.

Powered ageing assessments are also used to monitor thermal behaviour, power stability and overall system health before final inspection and shipment.

Together, these processes are intended to improve the readiness of systems before they reach customer facilities, where installation and integration can involve significant infrastructure requirements.

Storage and Networking Complete the AI Infrastructure Stack

ASUS is extending its SC26 presentation beyond GPU computing to include storage and data infrastructure, which are essential for supporting large AI workloads.

The company is showcasing the VS320D-RS26N enterprise storage platform and the OJ340A-RS60 high-density object storage system. It is also presenting the RS500A-E14B-R12U, a 1U server powered by sixth-generation AMD EPYC 9006 processors, with PCIe 6.0 connectivity and NVMe storage.

These systems address different requirements around storage capacity, data access and computing resources. As AI environments expand, organisations need infrastructure that can move large volumes of data efficiently between storage systems and accelerated computing platforms.

The broader portfolio reflects ASUS’s effort to position itself as an infrastructure provider rather than simply a supplier of individual server components.

Digital Twins and Deployment Services Extend the AI Factory Model

A major part of the ASUS presentation is its AI Factory planning and deployment framework, which combines system engineering with NVIDIA’s DSX reference design and Omniverse DSX Blueprint for AI Factory Digital Twins.

The approach allows organisations to simulate facility requirements and system configurations before physical infrastructure is installed. ASUS is working with ecosystem partners including Schneider Electric, AVEVA and IBM to bring IT infrastructure, power, cooling and facility models into a high-fidelity simulation environment.

Digital-twin simulations can help identify potential constraints, compare design options and assess infrastructure requirements before construction or installation begins. This is particularly relevant for AI facilities, where power availability and cooling capacity can influence the scale and configuration of deployments.

ASUS Professional Services support the process through site-readiness assessments, power and cooling planning, system integration and on-site installation coordination.

After the hardware is installed, the ASUS Infrastructure Deployment Center automates provisioning, cluster configuration and software installation. ASUS Control Center Data Center Edition then provides centralised monitoring, security oversight and update management for ongoing operations.

This connects planning, simulation, deployment and lifecycle management within a single service framework.

AI Infrastructure Providers Focus on Delivery and Operational Readiness

The expansion of AI infrastructure is creating new demands on server manufacturers and system integrators. Large-scale AI deployments require close coordination between computing hardware, networking, storage, power systems and cooling infrastructure.

For customers, delays or design problems in any one of these areas can affect the readiness of an entire facility. This makes engineering validation, manufacturing consistency and deployment support increasingly important alongside processor and GPU performance.

ASUS’s latest announcement highlights how it is responding to these requirements through increased production of NVIDIA Vera Rubin NVL72 systems and a broader AI Factory offering.

The company says its servers are available worldwide, although individual product configurations may vary by region. It has not disclosed specific shipment volumes, customer order figures or pricing for the systems featured at SC26.

As demand for advanced AI computing continues to shape the data centre market, the ability to deliver integrated infrastructure at scale could become an important competitive factor. ASUS is seeking to address that requirement by combining rack-scale systems with manufacturing automation, digital-twin planning and ongoing operational services.

Leave a Reply

Your email address will not be published. Required fields are marked *