NVIDIA DGX Spark 64GB Gives Developers More Ways to Build and Scale Local AI
Local AI is becoming more useful by the token. As AI agents move from experiments into everyday development, increasingly capable open models are shrinking to fit on more devices, giving builders more to run locally. Coming this month, NVIDIA DGX Spark will be available with 64GB of unified memory from top manufacturer partners — Acer, […]
NVIDIA's DGX Spark 64GB, set to release this month, offers developers a new way to build and scale local AI. With 64GB of unified memory from partners like Acer, ASUS, Dell, Gigabyte, HP, and MSI, this compact AI supercomputer is designed for private use, eliminating cloud dependency. The platform comes with NVIDIA AI software stack, DGX OS, and NVIDIA Agent Toolkit, enabling developers to experiment with models and data without cloud intervention.
Two units can cluster together via NVIDIA Sync Cluster Assistant, doubling memory and expanding model support to up to 200 billion parameters. This personal AI supercomputer is accessible at an affordable price point, supporting models up to 100 billion parameters and agentic applications. DGX Spark is ready for agent development out of the box, with NVIDIA Agent Toolkit, CUDA-X AI libraries, Nemotron open models, and popular runtimes like Ollama, vLLM, and PyTorch with CUDA.
The platform also supports running Qwen 3.8 27B model on a single system or a cluster, simplifying AI development and scaling. NVIDIA Sync Cluster Assistant seamlessly configures multi-node clusters, detecting and validating devices while configuring the ConnectX-7 network for effortless scaling. Developers can start with a single unit for models that fit within its memory or scale to larger workloads by connecting multiple DGX Spark systems.
Written by urgent.news from NVIDIA Blog's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
