ASUS DGX Spark SSD Upgrade: Replace 1 TB1TB with 2 TB2TB and Reinstall DGX OS
Published 19/05/2026
Description
In this video, I upgrade my ASUS DGX Spark from a 1 TB1TB M.2 SSD to a 2 TB2TB M.2 SSD, then reinstall the official DGX OS from scratch. I show the full process: opening the device, replacing the internal 2242 NVMe SSD, downloading the ASUS / NVIDIA DGX OS image, creating a bootable USB installer, exploring the BIOS, and reinstalling Ubuntu-based DGX OS on the new drive. Contents: 0:00 - Introduction 0:17 - Description 1:50 - Open DGX 2:09 - Replace M.2 2:50 - Dig Deeper 3:52 - Close Back up 4:52 - Download OS 8:09 - Boot without OS 12:11 - Install DGX OS 16:54 - Disk space 17:08 - Update 17:38 - Conclusion The base ASUS DGX Spark model comes with a 1 TB1TB SSD, but for AI workloads, local LLM experiments, development, containers, models, and datasets, storage fills up quickly. Since this system only supports a single internal M.2 drive, upgrading the SSD means replacing the original disk and reinstalling the operating system. What you’ll learn in this video: Opening the ASUS DGX Spark chassis Identifying the internal 22422242 M.2 NVMe SSD Replacing the original 1 TB1TB drive with a 2 TB2TB SSD Looking at the internal design, cooling, fans, and heatsink Downloading the official ASUS DGX OS / NVIDIA DGX OS image Extracting the ISO from the downloaded archive Creating a bootable USB installer with Rufus Booting from USB and reviewing BIOS / UEFI settings Reinstalling DGX OS on the new SSD Verifying that the system detects the full 2 TB2TB capacity Updating the system after installation If you are using the DGX Spark as an AI workstation, developer box, inference node, or headless Linux machine, upgrading storage can make a big difference once models, containers, checkpoints, and datasets start piling up. If this helped you, don’t forget to like, subscribe, and leave a comment with your setup or questions. #DGXSpark #Ubuntu #NVMe Follow me on X: https://l.samayas.eu/jo3eaYEf3E?P=YT&V=127 Buy me a Coffee : https://