High-Performance Computing in a Small Form Factor

iBuyPower has officially entered the compact AI workstation market with its new offering: the Pro AMD AI Halo Mini PC. Priced at $3,999, this machine is engineered to deliver high-end professional performance and support local AI workloads while maintaining a minimal desk footprint of just 70 × 198 × 200 mm.


The system distinguishes itself with a unique design aesthetic, featuring a leather strap attached to the compact chassis, setting it apart from traditional utilitarian workstations.


Technical Specifications and AI Capabilities

At the core of the workstation lies AMD’s Ryzen AI Max+ PRO 495 processor. This 16-core Zen 5 chip is paired with integrated Radeon 8065S graphics based on RDNA 3.5 architecture, which boasts 40 compute units. For AI-specific acceleration, the integrated neural processing unit (NPU) delivers 55 TOPS. To support data-heavy operations, the system includes 128GB of high-speed LPDDR5X memory running at 8,000 MHz.


«iBuyPower indicates that the hardware is built to handle intensive creative, engineering, and data processing tasks, alongside local AI model execution, ray tracing, and interactive simulations.»

Connectivity and Expandability

The device is equipped with a robust set of connectivity options, including:

  • Networking: Wi-Fi 7, Bluetooth 5.4, and dual 10Gb Ethernet ports.
  • Front I/O: Two USB 3.2 Gen 2 Type-A ports, a USB4 Type-C port, an audio jack, and an integrated SD card reader.
  • Rear I/O: DisplayPort 1.4, HDMI 2.1, an additional USB4 port, and supplementary USB connectivity.

Regarding storage, the system provides three M.2 slots—two 2280 positions and one flexible slot compatible with 2242 or 2280 drives, all utilizing PCIe 4.0 bandwidth.


Availability and Configuration

While the base configuration starts at $3,999 with a 1TB SSD, iBuyPower has confirmed that customers will have the ability to select different storage capacities upon launch via their website. Although the inclusion of dual 10Gb Ethernet ports is notable, the company has not yet detailed specific software configurations for cluster-based AI workloads using multiple units. The system is currently slated for release in Q4 2026.