Do you always need the cloud to run AI applications? And what changes when your data can be processed on a device sitting right on your desk?
In this MENA TECH interview, we speak with Jackson K.L WAT from GIGABYTE about AI TOP ATOM, a compact desktop AI supercomputer designed to bring AI computing closer to developers, researchers, and students.
Wat explains the device’s relationship with NVIDIA’s DGX Spark platform and what distinguishes GIGABYTE’s implementation, from its chassis design to improvements in cooling and thermal management. He also discusses why these design choices matter under demanding workloads and how the compact system differs from larger, more powerful workstations.
A central theme of the conversation is the case for running AI locally. Wat discusses the benefits of keeping data on a local device, reducing reliance on external cloud services, and carrying out supported tasks without a continuous internet connection. He shares his perspective on why data privacy and control remain important considerations when choosing where AI workloads should run.
Who is this device built for? We explore its intended audience, including AI application developers, research engineers, university students, and professors. Wat also explains how access to local computing resources could reduce their dependence on recurring cloud subscriptions, highlighting the practical considerations behind choosing a local system for development and research.
Looking ahead, we ask how this category of devices could evolve over the next five years and whether future versions might become even smaller.
The conversation concludes with a discussion of RAM and flash storage prices. Could a wider shift toward local AI computing help bring costs down? Wat shares his outlook and explains why, at the time of the interview, he did not expect prices to decline in the short term.
Watch the full interview for GIGABYTE’s perspective on local AI computing, the role of compact AI systems, and what AI TOP ATOM could offer users working on AI projects.
For your projects, what matters most when choosing between local hardware and cloud services: data privacy, cost, or performance?
