GIGABYTE AI TOP system highlights shift to local AI

GIGABYTE is showcasing its AI TOP product line at CES 2026, underscoring how the rapid adoption of AI inferencing is accelerating the shift of AI workloads from cloud-based environments to local systems. According to the company, this transition is making AI more accessible, responsive, and adaptable for everyday use.

The approach aligns with the broader idea of enabling practical AI usage across a wide range of users and deployment scenarios.

Evolution of the AI TOP portfolio

Recognising the growing need for local AI development, GIGABYTE introduced the AI TOP system in 2024. The platform was designed to enable AI development using household-standard power and has since expanded to support diverse use cases, ranging from personal experimentation to business deployments.

At the core of the portfolio is the GIGABYTE AI TOP system, a fully customisable AI computing solution built on GIGABYTE AI TOP series hardware. The system is designed to match specific AI workload requirements and support local AI development on the desktop.

Configurations for different business needs

Within the lineup, the AI TOP 500 system targets medium-sized businesses seeking scalable local AI computing power. It supports AI models with up to 405 billion parameters and is positioned for more demanding workloads.

For smaller businesses, startups, and individual users, the AI TOP 100 system supports large language model fine-tuning of more than 110 billion parameters. To increase performance, multiple AI TOP systems can be clustered using Ethernet and Thunderbolt connections, enabling faster training and expanded computing capacity.

Introducing the AI TOP ATOM

GIGABYTE is also highlighting the AI TOP ATOM, described as a personal AI supercomputer designed for local AI prototyping, fine-tuning, and inference. The system delivers data centre-class AI performance in a compact desktop form factor.

Powered by the NVIDIA GB10 Grace Blackwell Superchip, AI TOP ATOM delivers up to 1 petaFLOP of FP4 AI performance and supports AI models up to 200 billion parameters. By clustering two AI TOP ATOM systems, model support can be extended to 405 billion parameters.

Paired with the NVIDIA AI software stack, the system is aimed at creators, researchers, and developers who require high-performance local AI computing with portability.

Expanding the local AI ecosystem

Beyond standalone systems, GIGABYTE is strengthening its local AI ecosystem through a broader portfolio that includes human-centred AI gaming laptops with a built-in AI agent, GiMATE, along with AI-enhanced components and scalable local AI solutions for high-performance workloads.

GIGABYTE is demonstrating its AI product lines at its booth during CES 2026, with additional media and invited sessions scheduled at dedicated venues.