ASUS has introduced a new wave of AI infrastructure solutions at the 2025 OCP Global Summit in San Jose, showcasing next-generation servers and AI systems designed to accelerate enterprise artificial intelligence adoption. Among the key highlights are the XA NB3I-E12 AI servers, the ASUS AI POD built on the NVIDIA GB300 NVL72 platform, and the compact Ascent GX10 personal AI supercomputer.
Delivering high-performance computing for enterprise AI
ASUS unveiled the XA NB3I-E12 series AI servers, built to meet the needs of enterprises and cloud service providers (CSPs) managing intensive AI workloads. Powered by the NVIDIA HGX B300 system, the servers integrate NVIDIA ConnectX-8 InfiniBand SuperNICs, five PCIe expansion slots, 32 DIMM, and 10 NVMe. This combination delivers high performance, stability, and scalability, unlocking the full potential of AI applications across industries.
Since September, ASUS has begun shipping both the XA NB3I-E12 servers and the ASUS AI POD built on the NVIDIA GB300 NVL72 platform, giving organisations early access to cutting-edge AI performance and reliability.
Accelerating adoption with the ASUS AI Factory
At the summit, ASUS also showcased its AI Factory, an end-to-end platform built on NVIDIA Blackwell architecture that simplifies and accelerates AI deployment for enterprises. The AI Factory brings together high-performance hardware, optimised software, and professional services into a unified solution that supports a broad range of applications, from generative AI and natural language processing to predictive analytics.
By integrating ASUS servers, rack-scale AI PODs, and high-serviceability designs, the AI Factory reduces deployment complexity, improves operational efficiency, and maximises computing resources. It enables enterprises to scale workloads from edge devices to large-scale AI supercomputing environments, helping them innovate faster and adapt to rapidly evolving AI needs.
On display at the summit, the AI Factory demonstrated how organisations can harness AI across diverse sectors, including manufacturing automation, smart cities, and data analytics. Its comprehensive ecosystem supports seamless integration, flexible deployment, and scalable infrastructure designed to evolve with future AI advancements.
Compact AI supercomputing with ASUS Ascent GX10
Expanding its AI solutions portfolio, ASUS introduced the Ascent GX10, a compact personal AI supercomputer available from 15 October. Powered by the NVIDIA GB10 Grace Blackwell Superchip, the GX10 delivers up to 1 petaFLOP of AI performance for intensive workloads. Equipped with an NVIDIA Blackwell GPU, a 20-core Arm CPU, and 128GB of memory, it supports AI models of up to 200 billion parameters, bringing petaflop-scale inferencing capabilities directly to developers’ desktops.
Optimising performance with AMD EPYC 9005 processors
ASUS also highlighted server solutions powered by AMD EPYC 9005 processors, designed for AI-driven, mission-critical data centre workloads. The ESC8000A-E13X server accelerates generative AI and large language model (LLM) applications and is fully compatible with the NVIDIA RTX PRO 6000 Blackwell Server Edition. Embedded NVIDIA ConnectX-8 SuperNICs support 400G InfiniBand or Ethernet per QSFP port, enabling ultra-low latency, high-bandwidth networking, and scalable performance with NVIDIA Quantum InfiniBand or Spectrum-X Ethernet platforms.
Additionally, the RS520QA-E13 series of multi-node servers are optimised for high-performance computing (HPC), electronic design automation (EDA), and cloud computing. Each node supports up to 20 DIMM slots with advanced CXL memory expansion, PCIe 5.0, and OCP 3.0, maximising efficiency and performance for demanding AI and data workloads.
With these new solutions, ASUS aims to strengthen the foundation for enterprise AI transformation. By combining advanced server technology, scalable AI infrastructure, and powerful personal supercomputing capabilities, ASUS is enabling organisations to accelerate innovation and harness the next generation of AI-driven opportunities.
