ASUS Introduces ProArt RTX Spark PCs for Local AI Workflows
The market for high-end creative workstations is shifting toward local artificial intelligence processing. ASUS has announced the ProArt P16, P14, and GR1X, a new lineup of Windows PCs designed to run complex generative models without relying on cloud services.
NVIDIA RTX Spark Architecture and Technical Specifications
These new systems utilize the NVIDIA RTX Spark architecture, which integrates Blackwell RTX GPUs with Grace CPUs. According to press.asus.com, this hardware combination supports up to 128GB of high-bandwidth unified memory. This shared memory pool allows the systems to handle large datasets and AI models that typically require server-grade hardware.
The technical capabilities of the RTX Spark platform include:
- 1 petaflop of AI performance for local computations.
- Fifth-generation Tensor Cores featuring FP4 support.
- NVIDIA NVLink-C2C interconnect technology for high-speed data transfer between components.
- 10GbE networking included on the compact GR1X model.
Portable Power vs. Compact Desktop Design
For mobile professionals, the ProArt P16 measures 12.9mm thin and weighs 1.77kg. It features a 16-inch 4K Lumina Pro OLED display. The smaller ProArt P14 reduces the footprint further with a 13.9mm thickness and a weight of 1.48kg, utilizing a 3K Lumina Pro OLED panel. What this means for you is the ability to run heavy AI workloads on a machine that fits in a standard backpack.
The ProArt GR1X serves as a desktop alternative, measuring 150 x 150 x 51mm. Despite its small size, ASUS includes dual-fan cooling to manage the thermal output of the Blackwell and Grace hardware. To support local-first visual AI, ASUS collaborated with Nous Research for the Hermes Agent and ComfyUI for workflow automation.
NVIDIA RTX Spark unites NVIDIA AI, RTX graphics, and unified memory in one slim, powerful Windows PC.
— press.asus.com
Software Integration and Practical Utility
ASUS preloads these machines with specific tools to utilize the 1 petaflop of performance. MuseTree handles local image generation, while StoryCube provides intelligent asset management for large media libraries. By keeping these processes local, you avoid subscription fees and latency issues associated with cloud-based AI tools.
Whether these machines maintain their performance under sustained thermal loads remains to be seen in real-world testing. For developers and creators needing 128GB of memory in a portable form factor, these units represent a significant shift in mobile workstation capabilities. Buy now if your workflow requires immediate, high-capacity local AI memory; otherwise, wait for thermal benchmarks of the slim chassis.