Giga Computing, a subsidiary of GIGABYTE Group, has announced worldwide availability of the GIGABYTE XL44-SX2-AAS1, an NVIDIA RTX PRO server designed to accelerate AI and data center workloads. This system uses NVIDIA RTX PRO 6000 Blackwell Server Edition graphics processing units, NVIDIA BlueField-3 data processing units, NVIDIA ConnectX-8 SuperNICs, and integrates with NVIDIA AI Enterprise software to enable unified compute, networking, and management for data-to-AI workflows.
The GIGABYTE XL44-SX2-AAS1 is based on the NVIDIA MGX modular reference design and supports up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition graphics processing units, providing up to 768 GB of GDDR7 memory. It offers dual Intel Xeon 6700 or 6500 series central processing units, 32 DDR5 memory slots, and 3+1 redundant 3,200 W 80 PLUS Titanium power supplies, which, according to Giga Computing, ensures high throughput for generative AI, scientific, and 3D rendering applications.
The server leverages the NVIDIA BlueField-3 data processing unit for 400 Gb/s of bandwidth to storage and external networks, significantly improving secure and efficient access. Four NVIDIA ConnectX-8 SuperNICs support up to 800 Gb/s of InfiniBand or Ethernet bandwidth per graphics processing unit, optimizing low-latency east-west GPU-to-GPU communication for large-scale distributed AI training environments commonly used in hyperscale and colocation data centers.
For software integration, the XL44 Series is coupled with NVIDIA AI Enterprise, including NVIDIA NIM microservices, Omniverse tools for digital twin simulation, and Cosmos for bridging virtual and real-world AI deployments. This setup aims to streamline model development, inference, and physical AI implementation for enterprise-scale users.
Giga Computing claims the XL44-SX2-AAS1 delivers up to five times faster AI inference compared to previous generations. The server supports a variety of operating environments, including Windows, Linux, Kubernetes, and virtualization platforms, to facilitate adoption in both new and existing data center architectures. Giga Computing identifies data center AI workloads as the primary target, but also lists use cases in robotic simulation, video analytics, predictive modeling, and 3D content creation.
The GIGABYTE XL44-SX2-AAS1 is now available globally. More details are provided by Giga Computing on its website.
Source: Giga Computing
