ASUS has unveiled a new approach to AI factories, moving beyond standalone hardware toward fully integrated, production-ready platforms. The announcement came at ASUS AI Tech 2026, where the company outlined a strategy that unites infrastructure, deployment and governance into a single continuous system.

From Systems to Full AI Factories

Previously, ASUS focused on building systems that simply run AI workloads. Now, the company is redefining its role. Instead, ASUS is positioning itself as a builder and operator of complete AI factories.

To achieve this, ASUS is working closely with NVIDIA. In addition, it has partnered with a growing list of technology companies, including Aleria, AMD, Crusoe, Foxlink, IBM, Intel, Samsung, Schneider Electric and WD.

Together, these partners aim to help enterprises engineer performance, resilience and governance across the entire AI lifecycle. As a result, organizations can shorten the time it takes to generate a first token. Furthermore, they can accelerate time to revenue and improve infrastructure efficiency, all while maintaining trusted AI at scale.

Executive Perspective

Paul Ju, Senior Vice President and Head of ASUS’s Infrastructure Solution Business Group, explained the thinking behind the shift.

According to Ju, ASUS is fully embracing its “All in on AI” initiative. This strategy embeds AI across infrastructure, commercial solutions and physical devices. He noted that as AI factories become always-on production environments, infrastructure performance and governance can no longer be treated separately. Consequently, ASUS is bringing together compute, connectivity, power, cooling and automation to convert resources into AI output.

Governance Starts Before the First Token

Trust in AI, ASUS argues, must be built before deployment even begins. Therefore, the company combines advanced system engineering with the NVIDIA DSX Sim Blueprint. This allows enterprises to design, simulate, validate and optimize AI factories before any physical installation takes place.

Working alongside Schneider Electric, AVEVA and IBM, ASUS merges compute, networking, storage, power, cooling and facility infrastructure into one digital twin environment. This, in turn, helps organizations spot constraints and optimize configurations early, reducing deployment risk before installation.

Once systems go live, governance does not stop. Instead, it extends across ongoing operations. Since agentic and physical AI are becoming more autonomous, enterprises need continuous visibility and control over models, data, tools and enterprise systems.

To support this, ASUS offers two infrastructure management platforms: ACC (ASUS Control Center) and AIDC (ASUS Infrastructure Deployment Center). Beyond these, the company’s AI Software Stack and Turnkey AI Applications extend governance further. Within the stack, customers can access a Quota & Billing tool and an MLOps Portal, both designed to streamline resource management and speed up operations.

At the application level, the AI Software Stack functions as an “operational harness.” It links enterprise governance directly to AI services and autonomous agents, enabling secure and production-ready deployment of agentic AI.

Product Lineup Across the AI Factory Spectrum

ASUS also detailed hardware built specifically for different AI factory use cases:

• Large-scale AI factories: The ASUS AI POD, XA VR721-E3, is built on NVIDIA Vera Rubin NVL72. It offers a rack-scale platform for high-volume intelligence and low-cost token generation, delivering 10 times the performance per watt compared to the previous generation. Meanwhile, the enterprise-ready XA NR1I-E12LR and XA NR1I-E12L, based on NVIDIA HGX Rubin NVL8, accelerate high-density infrastructure for training, inference and post-training workloads.
• Agentic AI: The ASUS XA P2N-E2 is a 2U NVIDIA MGX platform. It is equipped with two NVIDIA Vera CPUs and supports up to two dual-slot NVIDIA GPUs, making it suitable for responsive, energy-efficient reasoning and orchestration at scale.
• Edge and visual computing: The ASUS ESC8000-E12P features NVIDIA RTX PRO 6000 and RTX PRO 4500 Blackwell Server Edition GPUs. It extends acceleration for inference and vision AI from the data center to the edge. Additionally, the compact PE3000N edge-AI platform, powered by the NVIDIA Jetson Thor T5000 module, delivers real-time inference and autonomous control for robotics and industrial automation.
• Data foundation: The ASUS UF920-E3-RS24 server, built on the NVIDIA STX modular foundation, works alongside the OJ340A-RS60 object storage solution and VS320D-RS26N storage system. Together, they provide the resilient data foundation enterprise AI pipelines require, from design validation through production-scale deployment.

Real-World Applications on Display

Beyond hardware, ASUS showcased how partners are already applying this technology. Foxlink demonstrated advanced industrial automation using autonomous robotics. Meanwhile, Aleria highlighted the growing need for Sovereign AI deployments that support localized data control and secure operations.

By combining software-hardware integration with Kubernetes platforms, ASUS says it can significantly shorten time-to-revenue for customers. Rather than simply selling computing power, the company positions itself as a provider of operational AI factories designed to generate ongoing business value.

Looking Ahead

Going forward, ASUS plans to deepen its collaboration with NVIDIA and its wider partner ecosystem. By integrating accelerated computing, simulation-validated deployment and AI harness engineering into one governance framework, the company aims to help organizations protect infrastructure investments as AI systems grow more complex and autonomous.

Ultimately, ASUS says its goal is to help enterprises move from simply investing in AI to embedding intelligence throughout their operations. This shift is intended to shorten time to revenue, improve token-per-watt efficiency, reduce waste and total cost of ownership, and build AI factories engineered for performance, adaptability and scalable growth.

Availability and Pricing

ASUS servers are available worldwide. However, availability of certain other ASUS products depends on local regulatory requirements. Customers should contact their local ASUS representative for specific product availability and pricing in their region.

Related post

View all