lucadelladora – Nvidia is launching its DGX Spark, a mini PC designed specifically for AI developers, researchers, and students. The device weighs just 2.6 pounds and delivers data-center-level performance in a compact form. Nvidia began accepting preorders earlier this year, and the DGX Spark will officially go on sale on October 15 through Nvidia.com and select retailers like Micro Center.
Read More : Qualcomm Faces Chinese Scrutiny Over Autotalks Deal
Unlike traditional consumer desktops, the DGX Spark runs Nvidia’s custom DGX OS, a version of Ubuntu Linux tailored with AI software. This system optimizes the machine for AI workloads rather than general computing tasks. The device features Nvidia’s GB10 super chip, which combines a 20-core Arm-based Grace CPU with a Blackwell GPU. The GPU matches the CUDA core count found in the RTX 5070 graphics card.
Nvidia developed the DGX Spark to offer local, powerful AI training capabilities without requiring access to large data centers. The compact design fits easily on a desk, yet packs impressive hardware. It includes 128GB of LPDDR5x system memory shared between the CPU and GPU, plus 4TB of NVMe storage. For connectivity, the device offers four USB-C ports, Wi-Fi 7, and an HDMI output.
The DGX Spark runs on standard electrical power and claims to be the world’s smallest AI supercomputer. Its size and performance mark a significant step toward democratizing access to advanced AI computing. Pricing for the DGX Spark starts at $3,999, excluding taxes and tariffs. Though expensive for casual users, this device targets professionals who need high performance in a portable package.
Wider Ecosystem and Future Nvidia AI Hardware Plans
At Computex, multiple vendors displayed their own versions of the DGX Spark. Companies like Asus, Dell, Gigabyte, HP, Lenovo, and MSI plan to release variants powered by Nvidia’s GB10 chip. Acer will launch its Veriton GN100 AI Mini Workstation in December, demonstrating the growing ecosystem around Nvidia’s AI hardware.
Nvidia is also developing larger AI-focused machines. The upcoming DGX Station desktop tower will feature the more powerful GB300 Grace Blackwell Ultra chip. While Nvidia has not yet announced pricing for the DGX Station, it expects to release it later this year. Partners like Asus, Boxx, Dell, HP, and Supermicro will assist in bringing the product to market.
Read More : Vivo Unveils TWS 5 With Hi-Res Sound and Long Battery
Beyond the DGX Spark and DGX Station, Nvidia reportedly works on a Windows-based AI PC. This future device could combine Nvidia’s GPU technology with an Arm chip from MediaTek. This move suggests Nvidia’s broader ambition to integrate AI computing into diverse hardware platforms for developers and enthusiasts.
Overall, Nvidia’s DGX Spark represents a key advancement in AI hardware accessibility. Its compact size and powerful specs make it ideal for local AI model training. As AI technology evolves, Nvidia and its partners will likely expand this product line, pushing performance and portability further. Developers should watch for more innovations and new offerings in AI computing throughout 2025 and beyond.
