Gigabyte Pitches Ai Factories, Physical Ai And Agentic PCs At CES 2026

Gigabyte used its CES 2026 presence to expand its Ai ambitions, unveiling a full-stack portfolio under the theme “Ai Forward.” The Taiwanese vendor showcased Ai-focused data centre platforms, compact edge devices and consumer desktops aimed at local large language model (LLM) use cases.

The announcements continue Gigabyte’s shift from component maker to full-stack infrastructure provider, with new offerings covering hyperscale inference, warehouse robotics and agentic desktop computing.

Gigapod Targets Enterprise Ai Factories

At the centre of Gigabyte’s data centre pitch is Gigapod, a modular system for building out Ai infrastructure using what the company calls a “building-block design.” Gigapod bundles liquid-cooled servers, networking and Gigabyte’s own rack management software, GPM.

The G4L3 and G4L4 servers at the heart of the platform support Intel Xeon 6 and AMD EPYC 9004 or 9005 processors, depending on configuration. They are paired with either Nvidia HGX B300 systems or AMD Instinct MI355X accelerators.

Gigabyte claims the design delivers high compute density with better thermal efficiency. A new 1U rack management switch supports up to eight liquid-cooled racks and can communicate with multi-vendor CDU protocols while detecting leaks.

On the high end, the company unveiled a rack-level configuration using the Nvidia Grace Blackwell Ultra NVL72. This system includes 72 Nvidia Grace CPUs and is tied together using Nvidia Quantum-X800 InfiniBand and Spectrum-X Ethernet.

Gigabyte said this platform offers up to 50 times the inference performance of Nvidia’s previous Hopper-based generation.

For broader training and simulation workloads, the company also displayed two new systems: the G894-SD3-AAX7 powered by HGX B300, and the XL44-SX2-AAS1 featuring RTX PRO 6000 Blackwell Server Edition GPUs. Both integrate Intel Xeon 6 CPUs and Nvidia BlueField-3 DPUs.

Workstation users were not left out, with the W775-V10-L01 offering server-grade GPU performance and closed-loop liquid cooling in a desktop tower format.

Bringing Ai To The Edge

Gigabyte also pitched Ai at the edge, with a showcase focused on warehouse automation. The company demonstrated compact Ai inference computers, embedded systems for managing autonomous mobile robots, and industrial PCs for robotics and conveyor systems.

These systems are aimed at low-latency, always-on operations in environments where centralised cloud inference is either impractical or too slow.

By integrating Ai into operational technology, Gigabyte is positioning itself to support the rise of what it refers to as “Physical Ai” — decision-making and actuation at the source of data.

Agentic PCs For LLM And RAG Workloads

On the consumer side, Gigabyte introduced a new series of Ai desktops under the “Ai TOP” brand, designed to run LLM and RAG (retrieval-augmented generation) applications locally.

The Ai TOP ATOM, Ai TOP 100 Z890 and Ai TOP 500 TRX50 are intended for users seeking to fine-tune or operate models without relying on the cloud.

These desktops are designed for standard power sockets and offer secure offline processing. Gigabyte’s Ai TOP Utility software is bundled with the machines and provides a visual interface for model management and inference workflows.

The company also showed off new laptops with built-in Ai companions branded as GiMATE, alongside the AORUS RTX 5090 Ai BOX. The latter is an external GPU box with Thunderbolt 5 connectivity, designed to give notebooks access to the full performance of a GeForce RTX 5090 desktop GPU.

Expanding Ai Infrastructure Ambitions

While Gigabyte is better known for its motherboards and graphics cards, the company has invested heavily in enterprise systems over the past five years, particularly in Ai servers.

The CES display reinforced that direction, combining off-the-shelf hardware with vertically integrated software to offer what it calls “Ai Factories.”

With Nvidia Grace Blackwell platforms just entering the market and interest in localised agentic Ai on the rise, Gigabyte is positioning itself across both cloud and edge to support emerging workloads at scale.

Last Updated on January 6, 2026 by Nick Ross