Build Future-Ready AI Factory

GIGABYTE introduces its one-stop AI data center solution with GIGAPOD, a modular “Building-Block Design” that integrates high-performance servers, high-speed networking, and GIGABYTE POD Manager (GPM) software, streamlining AI infrastructure design, deployment, and validation—accelerating the creation of enterprise AI Factories.

At the core, GIGAPOD deploys G4L4/G4L3 direct-liquid-cooling servers supporting Intel® Xeon® 6 processors and based on NVIDIA HGX B300 systems, and AMD EPYC™ 9005/9004 processors with AMD Instinct™ MI355X accelerators. These servers balance extreme compute density with advanced thermal efficiency. In addition, GIGABYTE introduces an in-house Rack Management Switch that centralizes management for up to eight DLC racks within a compact 1U form factor. It supports multi-vendor CDU communication protocols and precise leak detection, significantly enhancing system reliability and management efficiency—establishing a new benchmark for intelligent, scalable AI data centers.

GIGABYTE further expands its server families for data centers at all scales. The flagship NVIDIA Grace Blackwell Ultra NVL72—a rack-level compute node with 72 NVIDIA Grace CPUs, complemented by NVIDIA Quantum-X800 InfiniBand and NVIDIA Spectrum-X Ethernet networking, can deliver up to 50× inference performance compared with the previous NVIDIA Hopper generation.

For training, simulation and high-volume inference, GIGABYTE presents the G894-SD3-AAX7 and XL44-SX2-AAS1 supercomputers, accelerated by NVIDIA HGX B300 and NVIDIA RTX PRO™ 6000 Blackwell Server Edition GPUs respectively. Both are equipped with dual Intel® Xeon® 6 processors, DDR5 memory, and high-speed InfiniBand/Ethernet connectivity, while integrating NVIDIA BlueField-3 DPUs to enhance data security and computational efficiency. The W775-V10-L01 workstation brings server-class GPU performance and closed-loop liquid cooling to creators and small enterprises deploying on-prem AI workflows.

Compact Edge Solutions for Physical AI

GIGABYTE extends AI from data centers to real-world operations through its embedded systems and industrial PCs—built for environments where low latency, reliability, and 24/7 edge inference are critical. At CES, this vision takes shape through a smart warehouse showcase featuring:

·compact edge computers for high TOPS AI inference,
·low-power embedded systems for AGV/AMR fleet coordination,
·industrial PCs for robotic arms and conveyor systems, and
·versatile designs with rich I/O for sensors and machine vision.

These solutions enable AI to think and act where data is created, marking a key step toward Physical AI.

Shape Everyday AI Agent

As the world enters the era of Agentic AI, GIGABYTE introduces the AI TOP series, the AI TOP ATOM, AI TOP 100 Z890, and AI TOP 500 TRX50, premium AI-empowered desktop computers that enable local LLM/LMM inference, fine-tuning, and RAG applications using standard electrical systems, ensuring private, secure, and affordable computing. Beyond exceptional hardware, GIGABYTE also introduces the AI TOP Utility software, which simplifies setup, offers an intuitive interface, and streamlines AI workflows for model management and deployment.

GIGABYTE also expands AI-optimized client computing with laptops integrated with the GiMATE AI companion, delivering on-device assistance for creators and professionals. The new AORUS RTX 5090 AI BOX, featuring Thunderbolt™ 5 connectivity and the GeForce RTX™ 5090 GPU, enables notebook users to access near-desktop AI and graphics performance.