The Pulse
Gigabyte says 64GB AI TOP ATOM desktop is now available
Gigabyte says its 64GB AI TOP ATOM is available alongside the existing 128GB configuration. The desktop comes with a choice of 1TB or 2TB SSDs, though availability and pricing vary by region.

AI.info Team · Written with AI · fact-checked ·
Gigabyte says the 64GB version of its AI TOP ATOM desktop is now available alongside the existing 128GB model. The November 11 announcement, made in German, puts another memory option in the lineup for local AI development.
The 64GB configuration is offered with either a 1TB or 2TB NVMe SSD in the M.2 format, using a PCIe 4.0 connection. Availability, sales channels, prices and storage configurations may vary by region.
Configurations and timing
Gigabyte had previously announced the 64GB model on October 2. That announcement put the start of availability at October 23, 2026. The later release confirms the company’s availability announcement and specifies the two storage choices.
Both memory configurations use the NVIDIA DGX Spark platform and the same compact case design. The system is intended for local AI development, with users able to run models and process project data in their own IT environments.
The company lists local model inference, prototype development, data analysis and application validation among the tasks supported by the desktop. It identifies offices, laboratories and educational settings as places where the system can be used.
Hardware and regional availability
The AORUS Official Store lists the 2TB version with an NVIDIA GB10 processor and 20 processor cores. That listed configuration has 64GB of LPDDR5x memory and a 2TB PCIe 4.0 NVMe SSD.
The store lists Wi-Fi 7, Bluetooth 5.4, one HDMI port and four USB 3.2 Gen 2x2 Type-C ports. The operating system is NVIDIA DGX OS.
The official store lists the 2TB configuration at €5,042.01 including VAT, with an expected release date of December 24, 2026. That listing is distinct from the company’s general availability announcement, which says regional offerings can differ.
Clustering and software
For supported distributed AI workloads, up to four AI TOP ATOM systems can be connected using built-in NVIDIA ConnectX-7 networking and the Cluster Assistant in NVIDIA Sync. Additional memory and computing resources depend on the application and software configuration.
In an earlier scientific computing demonstration, Gigabyte combined NVIDIA Nemotron open models with the NVIDIA NemoClaw open agent blueprint to connect research hypothesis generation with simulation workflows. The company presented the demonstration as an exploration of AI agents’ potential in scientific research.
The desktop combines the NVIDIA CUDA-accelerated AI software ecosystem with Gigabyte’s AI TOP Utility. The utility offers downloads of supported models, inference and retrieval-augmented generation, or RAG, for local workflows.
Gigabyte says developers can use those functions to try supported open models, test AI assistants and build knowledge-based question-and-answer applications using their own documents. The software is intended to support development, testing and validation of AI applications locally.
Sources
Related stories
- Epoch AI introduces index for measuring inference capacity
- OpenAI Says Ex-Employee Sent 100+ Device Files to Personal Accounts
- Nvidia’s Board Opens Independent Inquiry Over Reports on Jensen Huang
- Nvidia Sets a 51% Non-China Sales Threshold for Cloud Clients
- U.S. Probe Put Nvidia’s Jensen Huang Near an Alleged China Spy Case