TLDR: Giga Computing, a subsidiary of GIGABYTE Group, has announced the global availability of its new GIGABYTE XL44-SX2-AAS1 server. This NVIDIA RTX PRO Server is engineered to handle demanding AI and data workloads, integrating cutting-edge NVIDIA technologies like RTX PRO 6000 Blackwell Server Edition GPUs, BlueField-3 DPU, and ConnectX-8 SuperNICs, alongside the NVIDIA AI Enterprise software suite. The system aims to transform traditional data centers into advanced AI factories.
TAIPEI, November 5, 2025 – Giga Computing, a prominent subsidiary of GIGABYTE Group, today announced the worldwide availability of its groundbreaking GIGABYTE XL44-SX2-AAS1 server. This NVIDIA RTX PRO Server is specifically designed to meet the escalating demands of modern AI and data workloads, unifying compute, networking, and software layers to accelerate the entire data-to-AI workflow.
Unifying Compute and Networking Power
At the core of the XL44-SX2-AAS1’s exceptional performance is its robust hardware configuration. Built on the NVIDIA MGX modular reference design, the server supports up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Each GPU is equipped with 96GB of GDDR7 memory, culminating in a total of 768GB of GDDR7 memory across the system, ensuring exceptional throughput for generative AI, 3D rendering, and scientific workloads.
Processing power is further bolstered by dual Intel® Xeon® 6700/6500 CPUs and 32x DDR5 DIMMs. The system’s reliability is guaranteed by 3+1 redundant 3200W 80 PLUS Titanium power supplies.
For high-speed data transfer and secure networking, the XL44-SX2-AAS1 integrates an NVIDIA BlueField-3 DPU, providing a total of 400Gb/s bandwidth for secure and efficient data access to storage and external networks. Additionally, four NVIDIA ConnectX-8 SuperNICs support the eight GPUs with up to 800Gb/s NVIDIA InfiniBand or Ethernet bandwidth per GPU. This advanced networking design facilitates low-latency, high-bandwidth east-west GPU-to-GPU traffic, crucial for scalable distributed AI training.
Seamless Integration with NVIDIA AI Enterprise
The XL44 Series server is not just a hardware powerhouse; it is deeply integrated with the NVIDIA AI Enterprise software suite. This comprehensive ecosystem includes support for NVIDIA NIM microservices, NVIDIA Omniverse libraries and technologies for digital twins, and NVIDIA Cosmos world foundation models. This integration allows enterprises to scale their operations from initial AI model development to inference and physical AI deployments within a consistent and optimized environment.
From Data Centers to AI Factories
Also Read:
- Hitachi Vantara Introduces iQ Studio: A New Platform to Streamline Enterprise AI Agent Development and Governance
- Microsoft Unveils $110 Billion Global Investment in Advanced AI Data Centers
The XL44-SX2-AAS1 is poised to deliver up to 5X faster AI inference performance compared to the previous generation. Its versatility enables it to support a broad spectrum of AI workloads, including intelligent agents, large language model (LLM) inference, 3D rendering, 3D simulation, robotic simulation in smart manufacturing, financial decision modeling, AI-assisted medical research, and video analytics. The server’s compatibility with Windows, Linux, Kubernetes, and various virtualization environments ensures seamless deployment into existing x86 infrastructures. Giga Computing emphasizes that the XL44-SX2-AAS1 is more than a GPU server; it is the foundational technology for transforming data centers into full-scale AI factories, driving innovation across diverse industries.
