GUNNIR Showcases Dual Intel Arc Pro B70 Liquid-Cooled Workstation for Local AI

GUNNIR is bringing dual-GPU workstation solutions to market, offering a path for users who need high-end AI compute without the bulk of traditional server racks. The company showcased new hardware built around Intel’s Arc Pro B70, targeting professionals who run local AI models on compact desks. This move addresses a growing demand for dense AI performance in smaller footprints. We see this as a practical option for developers and researchers who need power but lack rack space.

Compact dual-GPU systems target local AI developers without server racks

The centerpiece of the showcase is a dual-card liquid-cooled workstation and a mini workstation variant. Both systems rely on the Intel Core Ultra 200HX 'Arrow Lake HX' processor to handle system tasks. This platform choice allows GUNNIR to pack significant graphics capability into a small chassis. The dual-configuration Arc Pro B70 setup is designed to maximize throughput for demanding workloads.

Specifications

• GPU: Intel Arc Pro B70 (Dual-card configuration)
• CPU: Intel Core Ultra 200HX "Arrow Lake HX"
• Cooling: Liquid cooling and Vapor Chamber
• Form Factor: Workstation and Mini Workstation

To manage the heat from two high-performance GPUs, GUNNIR employs liquid cooling and vapor chamber solutions. These thermal designs are necessary to maintain stability in the compact form factors offered. The company also exhibited MXM form-factor discrete GPUs based on the Arc Pro B50 and B65. This indicates a broader strategy to support various integration needs beyond just the B70 workstation.

GUNNIR states that these systems provide significant VRAM and AI compute power for local AI applications within a compact form factor. The devices are specifically targeted at local AI application scenarios. We have previously covered the Arc Pro B70 in the context of its benchmark performance against other high-end cards. See our earlier coverage on Intel Arc Pro B70 Beats RTX for more context on the chip's capabilities. Specific VRAM capacities and exact AI compute metrics in TOPS remain undisclosed in the current reports.

The hardware is currently available as a showcase model with no pricing or release dates confirmed. GUNNIR has not yet provided specific costs or availability windows for the dual-card workstation or the mini variant. Buyers interested in these compact AI workstations will need to wait for official commercial details. The focus remains on demonstrating the thermal and spatial efficiency of the dual-GPU design.