NVIDIA RTX Spark Compact Desktops Now Available, Built for Local AI Workloads
- NVIDIA's RTX Spark compact desktops are designed for local AI workloads, featuring a powerful superchip with a 20-core CPU and Blackwell GPU.
- The systems support up to 128GB of unified memory and can run demanding AI tasks, including large language models and real-time video editing.
- NVIDIA is expanding software support for RTX Spark, enhancing creative tools like Photoshop and Premiere with AI-driven features and GPU acceleration. Read more
NVIDIA has announced the availability of its RTX Spark compact desktops, bringing its Blackwell-powered AI platform to small-form-factor Windows PCs designed for developers, creators, and AI enthusiasts.
The first systems are now available from ASUS, Dell, and MSI, with additional models from Acer, GIGABYTE, and HP coming soon. Microsoft is also preparing its Surface RTX Spark Dev Box.
Unlike traditional desktops focused primarily on gaming or productivity, RTX Spark systems are designed to handle demanding local AI workloads, including running AI agents, generating images and videos, and fine-tuning AI models without relying entirely on cloud services.
Blackwell Graphics and Up to 128GB of Unified Memory
At the heart of RTX Spark desktops is a superchip combining a 20-core NVIDIA Grace CPU and an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores.
The systems support up to 128GB of unified memory and deliver up to 1 petaflop of AI performance. They also feature 10-gigabit Ethernet for transferring large datasets, model files, and creative assets across networked storage.
According to NVIDIA, the hardware can handle several demanding workloads, including running 120-billion-parameter large language models locally, rendering 3D scenes exceeding 90GB in real time, and editing 4:2:2 video in real time.
Running AI models locally can also help keep files and code on the user’s own machine while avoiding per-token cloud API charges for supported tasks.
More AI and Creative Software Support

NVIDIA is also expanding software support for RTX Spark systems, covering both development and creative workflows.
Developers can use native CUDA support and Linux development tools through Windows Subsystem for Linux. The platform also supports local AI agents such as Hermes Agent, with NVIDIA NemoClaw blueprints and Unsloth available for supported model fine-tuning workflows.
For creators, NVIDIA highlighted integrations between Photoshop and ComfyUI through Hermes Agent, allowing users to generate images and animations through natural-language instructions.
The company is also expanding support for Adobe’s creative applications. Premiere gains TensorRT-RTX acceleration for features such as Enhance Speech, Object Mask, Scene Edit Detection, and Auto Reframe. NVIDIA says the improvements can deliver more than 50% quicker turnaround on everyday H.264 video deliverables.
Photoshop also gains GPU acceleration for its Remove Tool and native Camera Raw support, while Substance 3D Painter 12.2 is now optimized for RTX Spark in public beta. Illustrator and InDesign are also getting native Windows on Arm versions, with Illustrator offering up to 30% higher performance compared to its emulated version.
Availability

NVIDIA RTX Spark compact desktops are now available from ASUS, Dell, and MSI. Systems from Acer, GIGABYTE, and HP are expected to arrive soon, alongside Microsoft’s Surface RTX Spark Dev Box.
NVIDIA has yet to provide specific pricing details for the newly announced compact desktop models in its announcement. Availability and configurations will depend on the respective manufacturers and markets.
With RTX Spark, NVIDIA is looking to make local AI computing more accessible through compact Windows systems that can handle AI development, content creation, and other demanding workloads without requiring a traditional workstation setup.
