PEGATRON Unveils Next Generation AI Platforms Powered by NVIDIA Vera Rubin NVL72 and NVIDIA HGX Rubin NVL8 at GTC 2026

About this update from Pegatron Corporation

SAN JOSE, Calif., March 16, 2026 /PRNewswire/ -- PEGATRON, a globally recognized total server solutions provider, will showcase its latest AI infrastructure platforms at NVIDIA GTC 2026. Powered by NVIDIA Vera Rubin NVL72, NVIDIA HGX Rubin NVL8, and NVIDIA RTX PRO™ 6000 Blackwell Server Edition GPUs, these platforms deliver breakthrough performance, rack-scale integration, and liquid-cooled efficiency for hyperscale and enterprise AI deployments. NVIDIA Vera Rubin NVL72 Rack-Scale AI Supercomputer The Pegatron RA4803-72N3, powered by NVIDIA Vera Rubin NVL72, is a liquid-cooled rack-scale AI supercomputer designed for next-generation workloads. Integrating 72 NVIDIA Rubin GPUs, 36 NVIDIA Vera CPUs, NVIDIA NVLink 6 scale-up networking, NVIDIA ConnectX-9 SuperNICs, NVIDIA BlueField-4 DPUs, and NVIDIA Spectrum-X Ethernet networking, the solution delivers 3.6 EFLOPS inference performance, 260 TB/s bandwidth, and 20.7 TB HBM4 memory. Compared with NVIDIA Blackwell architecture, it trains mixture-of-experts models using one-fourth the GPUs, reduces inference costs by up to 10X, and achieves 5X higher throughput—dramatically improving performance per watt, lowering footprint, and optimizing total cost of ownership for AI factories. Powering Rack-Scale AI with Up to 9 HGX Rubin NVL8 Systems The Pegatron AS210-2T1-8H3 is a 2U liquid-cooled server based on NVIDIA HGX Rubin NVL8 and dual Intel Xeon 6 processors, designed to accelerate enterprise AI workloads. Leveraging NVIDIA Rubin GPUs, NVLink 6, ConnectX-9 SuperNICs, BlueField-4 DPUs, and Spectrum-X Ethernet networking, it delivers optimized AI and HPC performance. Scaling up to rack-level AI infrastructure, the RA4800-64H3 and RA4800-72H3 deliver ultra-high-density performance in liquid-cooled rack architectures. The RA4800-64H3 integrates eight 2U systems per rack, supporting up to 64 NVIDIA HGX Rubin GPUs and 16 Intel Xeon 6 processors, while the RA4800-72H3 integrates up to nine 2U systems in a 48U rack—scaling to 72 GPUs and 18 CPUs to enable unprecedented compute density and faster time-to-insights for large-scale AI workloads. NVIDIA RTX PRO 6000 and NVIDIA RTX PRO™ 4500 Blackwell Server Edition GPUs for AI and Digital Workloads Pegatron's servers powered by NVIDIA RTX PRO Blackwell Server Edition GPUs enable generative AI, data processing, simulation, and visualization workloads with exc...

View stock analysis, news, and events for Pegatron Corporation