ASUS launches ProArt GR1X mini PC with 128GB memory for local AI
The new compact workstation uses NVIDIA's Blackwell architecture to run large language models locally without relying on the cloud.
The short version
- Supports up to 128 GB of unified memory for running large language models (LLMs) locally.
- Powered by a 20-core Grace CPU and a 6,144-core Blackwell RTX GPU.
- Compact chassis measures 150 x 150 x 51 mm and weighs only 1.48 kg.
- Capable of handling models with up to 120 billion parameters and 1 million tokens of context.
Why it matters
This hardware allows professionals to run complex AI models on their own machines rather than sending sensitive data to the cloud. It bridges the gap between massive data center power and portable, desktop-friendly hardware.
A compact power house
ASUS has officially released the ProArt GR1X, a mini PC designed to handle high-performance computing tasks in a very small footprint. The device is built for users who need professional-grade power for AI workloads, gaming, and heavy CPU tasks without the bulk of a traditional tower. It is specifically engineered to provide a local alternative to cloud-based processing.
The hardware inside the ProArt GR1X is powered by NVIDIA's recently launched RTX Spark technology. This includes a 20-core Grace CPU and a 6,144-core Blackwell RTX GPU. By combining these components, ASUS aims to provide a platform that can manage intensive operations while remaining highly portable for various workspace setups.
Despite its internal power, the machine is remarkably small. It measures just 150 x 150 x 51 mm and weighs only 1.48 kg. This compact design makes it an ideal choice for users with space constraints who still require the ability to run demanding software. It is designed to function as a primary workstation for 24/7 AI operations or high-end gaming.
The unified memory advantage
One of the standout features of the ProArt GR1X is its unified memory design. Unified memory allows the CPU and GPU to share the same pool of memory, which is a critical requirement for running large language models (LLMs) efficiently. This architecture helps the machine process massive amounts of data without the bottlenecks found in traditional separate memory systems.
According to ASUS, this setup allows the ProArt GR1X to support models with up to 120 billion parameters. It can also handle up to 1 million tokens of context. For researchers and developers, this means they can run sophisticated AI models entirely on their own hardware. This local execution ensures that sensitive files and prompts remain on the user's machine rather than being sent over the internet.
While the machine is built for local use, it does not strictly forbid external connections. Users still have the flexibility to choose cloud or hybrid processing depending on their specific requirements. However, the primary goal of this hardware is to give enthusiasts and professionals the ability to keep their data private while maintaining high performance.
Cooling and connectivity
Managing the heat generated by high-performance components in a small chassis requires specialized engineering. The ProArt GR1X utilizes a dual-fan cooling solution to keep the hardware stable during intense workloads. This system includes 218 fan blades, 57 ultra-wide fins, and four heat pipes designed to handle a thermal design power (TDP) of up to 140W.
The device is packed with modern ports to ensure it can connect to a wide array of peripherals. It features USB 3.2 Gen 2x2 Type C ports, a 10 G LAN port, and WiFi 7 for high-speed internet. It also includes Bluetooth 5.4 and HDMI 2.1, which allows it to support up to four 4K displays or a single 8K display.
Storage options are also robust, featuring two M.2 NVMe slots that support both PCIe 5.0 and 4.0 interfaces. These high-speed slots ensure that data can be moved quickly between the storage drives and the processor. These specifications make the ProArt GR1X a versatile tool for both creative professionals and AI engineers.
Availability and pricing
ASUS is offering the ProArt GR1X in two primary variants to suit different needs. One version comes with 64 GB of unified memory, while the more powerful version includes 128 GB of unified memory. Both variants offer different storage configurations, with options for either 1 TB or 2 TB of storage space.
The official website lists the starting price at $4,499, though the final cost may vary depending on the region. The machine is being distributed through several major retail channels to reach a wide audience of enthusiasts. These include the ASUS eShop, Best Buy, Amazon, Micro Center, and Central Computers.
By placing this machine in major retail channels, ASUS is positioning the ProArt GR1X as a mainstream option for high-end local AI. It targets the growing market of users who want to experiment with LLMs and professional workloads without needing to build a custom server. It serves as a ready-to-use solution for those who want to skip the complexity of manual hardware assembly.
What to watch
- How the 120 billion parameter model performance compares to cloud-based alternatives.
- Whether more vendors will release similar RTX Spark-powered mini PCs soon.
Try it
Check the official ASUS website to see if the 128GB memory variant is available in your specific region.
Sources and links
- Officialofficial website eshop.asus.com
- ReportingWccftech: the article we sourced
First reported by Wccftech. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.