ASUS Prepares to Launch NVIDIA HGX Rubin NVL8 Systems for Next-Generation AI Workloads
ASUS has announced that its new server systems powered by NVIDIA HGX Rubin NVL8 are ready for launch in the fourth quarter of 2026. Available in both partial-rack and full-rack configurations, the systems are designed to meet the growing demand for high-performance AI infrastructure, particularly for advanced reasoning, large-scale model execution and other computationally intensive workloads.
The new portfolio includes the 3U ASUS XR R1I12B-M1BHFS hybrid-cooled server and the 2U ASUS XR R1I12B-M1BLFS, which features a fully liquid-cooled architecture. Both systems are designed to give enterprises greater flexibility when deploying AI infrastructure, allowing them to select configurations that align with their performance, cooling and data center requirements.
ASUS is showcasing the systems at NVIDIA GTC Berlin 2026, where visitors can explore the company’s broader range of AI computing solutions. The portfolio extends beyond data center servers to include desktop AI supercomputers, professional workstations, mini PCs and laptops designed for developers, researchers and creative professionals.
Together, these products reflect ASUS’s approach to supporting the complete AI development lifecycle, from large-scale infrastructure and model training to edge computing, inference and AI-assisted content creation.
Flexible NVIDIA HGX Rubin Systems for Enterprise AI
As AI models become more sophisticated and workloads grow increasingly demanding, enterprises require computing platforms that deliver high performance while managing power consumption, thermal output and deployment complexity.
ASUS’s new NVIDIA HGX Rubin NVL8 systems are designed to address these requirements through two distinct server configurations. By offering hybrid-cooled and fully liquid-cooled designs, the company aims to support different data center environments and infrastructure strategies.
The systems are intended for demanding AI applications, including large-scale AI reasoning and workloads that require substantial computing resources. Their flexible configurations also give organizations options when planning infrastructure expansion, whether they need individual server deployments or larger rack-scale installations.
ASUS XR R1I12B-M1BHFS: Hybrid Cooling in a 3U Design
The ASUS XR R1I12B-M1BHFS features a 3U form factor and a hybrid cooling architecture. It uses direct-to-chip liquid cooling for the NVIDIA HGX Rubin eight-GPU baseboard, while dual Intel Xeon 6 processors are supported by air cooling.
This approach combines liquid cooling for the high-performance GPU subsystem with air cooling for the CPU components, providing an alternative for facilities that require advanced thermal management but prefer to retain some conventional cooling infrastructure.
The system also features an air-cooled CPU sled designed to reduce maintenance time by up to 50%, according to ASUS. Faster maintenance can help simplify servicing and reduce the operational effort associated with keeping AI infrastructure running.
Flexible U.2 and E1.S storage options provide additional deployment choices, enabling organizations to configure storage according to their specific requirements.
The hybrid-cooled design may be particularly relevant for enterprises looking to introduce advanced GPU infrastructure while balancing cooling capabilities, serviceability and existing data center constraints.
ASUS XR R1I12B-M1BLFS: Fully Liquid-Cooled Performance
The ASUS XR R1I12B-M1BLFS takes a different approach with its 2U, fully liquid-cooled architecture. Designed for demanding AI environments, the system focuses on thermal efficiency, computing density and data center power usage effectiveness (PUE).
By applying liquid cooling throughout the system, the platform is designed to support high-performance computing while helping data centers manage the heat generated by intensive AI workloads.
The server also features an NVIDIA-validated design and a cable-less board-to-board architecture. ASUS says this approach helps simplify system integration and reduce deployment complexity, potentially making it easier for organizations to incorporate the platform into larger AI infrastructure environments.
The compact 2U design is intended for deployments where rack space and computing density are important considerations. Combined with full liquid cooling, it provides an option for organizations building infrastructure around the thermal and performance requirements of advanced AI computing.
The two systems give customers a choice between hybrid and fully liquid-cooled architectures, depending on their facility design, maintenance preferences and deployment goals.
Expanding the AI Portfolio Beyond Data Centers
At GTC Berlin 2026, ASUS is also demonstrating a broad portfolio of NVIDIA-powered systems covering AI development, professional computing and edge applications.
The ASUS Ascent GX10 is designed for AI model fine-tuning, prototyping and inference, offering a compact platform for developers working on AI applications. The ExpertCenter Pro ET900N G3 targets more demanding professional workloads, including AI development and large-scale model training.
For creators and users who need AI capabilities closer to their everyday workflows, ASUS is presenting the ProArt GR1X Mini PC and ProArt P16 laptop. These systems extend AI computing beyond traditional data center environments, supporting applications that benefit from local processing and specialized hardware.
The broader portfolio demonstrates how AI infrastructure needs vary across different users. Large data centers require high-density GPU systems, while developers and professional users may need accessible platforms for experimentation, application development and inference.
By offering products across these segments, ASUS aims to provide solutions for organizations at different stages of AI adoption, from early development to large-scale deployment.
Optimizing AI Factories With NVIDIA DSX Sim
Beyond server hardware, ASUS is highlighting its work on AI factory planning and simulation using the NVIDIA DSX Sim Blueprint.
Designing an AI factory involves coordinating multiple interconnected systems, including computing hardware, networking, storage, power delivery, cooling and facility infrastructure. Decisions made in one area can influence performance and efficiency throughout the installation.
ASUS combines its system engineering expertise with NVIDIA DSX Sim to help enterprises design, simulate and validate AI factory infrastructure within a digital twin environment. This approach enables organizations to evaluate potential configurations and identify constraints before committing to physical deployment.
Working with ecosystem partners Schneider Electric, AVEVA and IBM, ASUS is integrating computing, networking, storage, power, cooling and facility systems into a unified digital representation.
This environment is intended to help customers assess infrastructure layouts, investigate potential bottlenecks and optimize performance and efficiency ahead of implementation. Such simulation capabilities can be especially useful for large-scale AI facilities, where changes to power distribution, cooling capacity or equipment placement can have significant operational implications.
The goal is to help organizations make better-informed infrastructure decisions while preparing their facilities to support increasingly demanding AI workloads.
ASUS Showcases Its AI Solutions at GTC Berlin 2026
ASUS is inviting attendees to visit Booth #3039 at Station Berlin during GTC Berlin 2026 to explore its latest AI solutions across data center infrastructure, edge computing and professional applications.
The company is also highlighting its broader edge-to-cloud portfolio through the ASUS VIP Room at W3 Hub, which is scheduled to be open from October 20 to 22. Visitors can explore additional systems, including NVIDIA HGX-based platforms and enterprise infrastructure powered by NVIDIA RTX PRO technology.
With the planned fourth-quarter launch of its HGX Rubin NVL8 systems, ASUS is expanding its offerings for organizations preparing for the next generation of AI workloads.
The combination of hybrid-cooled and fully liquid-cooled servers, professional AI systems and digital twin-based infrastructure planning reflects the company’s focus on addressing multiple aspects of AI deployment.
As enterprises continue investing in advanced models and AI-powered applications, the ability to balance computing performance, thermal efficiency, maintenance and infrastructure planning will remain important. ASUS’s latest systems are designed to give organizations additional options for building AI environments capable of supporting those evolving requirements.
Source Link: https://press.asus.com/