ASUS has unveiled its latest range of AI infrastructure solutions at GTC Paris 2025, presented during VivaTech as a Gold Sponsor. The company also marked a new milestone in its ongoing partnership with Nebius, with both parties set to collaborate on the deployment of NVIDIA’s next-generation GB300 NVL72 platform. The announcement reflects ASUS’s strategic focus on scaling AI infrastructure for enterprises, from data centres to developer desktops.
Expanding collaboration with Nebius
ASUS is advancing its AI Factory vision with full support for NVIDIA Blackwell Ultra solutions. This includes strengthening its partnership with Nebius, a long-time collaborator in AI deployment. The two companies are now moving beyond the NVIDIA GB200 NVL72 platform and working together on the new GB300 NVL72 platform to support scalable, high-performance infrastructure.
Andrey Korolenko, Chief Product and Infrastructure Officer at Nebius, said: “We have collaborated with ASUS for many years and appreciate its impressive capability to deliver swift and efficient solutions. ASUS not only delivers consistently against our exacting technical requirements, but also demonstrates deep professional expertise in building AI infrastructure. The company’s forward-thinking approach and technical excellence have been a key enabler for our projects, and we look forward to working together to deliver the next generations of AI infrastructure.”
New AI infrastructure for data centres
At the event, ASUS introduced its latest AI Factory infrastructure built on NVIDIA RTX PRO Servers and Grace Blackwell Ultra systems. This includes the ASUS AI POD powered by the NVIDIA GB300 NVL72, designed for complex inference workloads. Another highlight is the ASUS XA NB3I-E12, which features the NVIDIA HGX B300 system and supports 2.3TB of HBM3e memory. This system is geared towards accelerating training and inference for large-scale AI models.
To support high-performance computing (HPC) environments, ASUS also revealed the ESC8000A-E13P server. This 4U NVIDIA MGX system can support up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, along with BlueField-3 DPUs and ConnectX-8 SuperNICs. With integrated PCIe 6.0 switches, the system is designed for seamless deployment in enterprise data centres, providing the flexibility and scale needed for modern IT operations.
Bringing AI power to the desktop
ASUS also debuted several solutions aimed at AI inferencing at scale. Among them is the ExpertCenter Pro ET900N G3, which is the first system to use the NVIDIA GB300 Grace Blackwell Ultra Superchip, offering up to 784GB of coherent memory. Complementing this is the ASUS Ascent GX10, a compact AI desktop supercomputer powered by the NVIDIA GB10 Grace Blackwell Superchip. The GX10 delivers 1,000 AI TOPS of performance, with a NVIDIA Blackwell GPU, a 20-core Arm CPU, and 128GB of memory.
Both systems support models with up to 200 billion parameters, bringing petaflop-class performance to the desktop. Optimised for training and inference workloads, they are designed for AI developers and teams that require high-end local compute capabilities. The integration of NVIDIA’s AI software stack further enables seamless deployment for research and enterprise needs.
ASUS’s AI infrastructure expertise
ASUS has a long history in server technology, dating back to 1995. It continues to provide end-to-end infrastructure solutions, from components to fully integrated systems, backed by global R&D and manufacturing. The company’s approach focuses on helping enterprises scale their AI initiatives through customisation, technical support, and a reliable track record of deployments. Under the banner of its “Ubiquitous AI, Incredible Possibilities” vision, ASUS remains committed to supporting the next wave of AI innovation.
