Pegatron will showcase NVIDIA Vera CPU liquid-cooled systems and NVIDIA Groq 3 LPX to meet evolving agentic AI infrastructure needs, with Vera CPUs optimized for tool calls, sandboxed code, and reinforcement-learning loops, deployed at rack scale alongside GPU infrastructure for efficient orchestration and data processing, and Groq 3 LPX enhancing latency-sensitive inference.

At the core is the RA4803-72N3 rack-scale AI platform, integrating 72 NVIDIA Rubin GPUs, 36 NVIDIA Vera CPUs, 6th generation NVIDIA NVLink, and fully liquid-cooled rack-level integration, demonstrating Pegatron’s ability to design and deploy complex rack-scale AI solutions.

AMD EPYC 9006 platforms are coupled with OCP 3.0, DC-MHS, and 48VDC busbar configurations to enable flexible, scalable rack architectures for compute and storage.

The presentation highlights end-to-end deployment considerations beyond compute performance, covering system-level engineering, liquid cooling, power architecture, open standards, manufacturing readiness, and rigorous validation using ORv3 busbar, DC-MHS, and OCP 3.0 network adapters.

Pegatron’s AI-ready compute and storage infrastructure includes AMD EPYC 9006 processors in platforms such as the SS201-1A1 storage server (2U, up to 32 E3.S NVMe SSDs) and GS203-2A2 and GS205-2A2E systems, offering up to 256 cores, 512 threads, high memory bandwidth, and PCIe Gen 6 connectivity.

Gary Cheng, Pegatron President and Co-CEO, emphasizes rapid movement from AI infrastructure design to production deployment by tightly integrating design, manufacturing, OCP-aligned architecture, and rack-level validation.

The AS210-2T1-8H3 system, a 2U fully liquid-cooled NVIDIA HGX Rubin NVL8 platform, supports configurations with dual Intel Xeon, AMD EPYC 9006, or NVIDIA Vera CPUs and eight Rubin GPUs, enabling high-density rack-scale options like RA4800-64H3 and RA4800-72H3.

Overall, Pegatron aims to move AI infrastructure from individual systems to production-ready clusters and data centers, translating concept and simulation into validated, scalable deployment.

Pegatron will showcase AI infrastructure capabilities at the OCP Global Summit 2026 under the theme Scaling Innovation for the AI Era, demonstrating how AI platforms translate into production-ready systems, racks, and clusters for data centers.

Pegatron is expanding capabilities through NVIDIA DSX AI factory workflows, linking digital-twin simulation with physical manufacturing, automated production, thermal validation, and run-in testing to convert engineering insights into production-ready AI infrastructure.