Giga Computing, a subsidiary of GIGABYTE, will exhibit a range of systems accelerated by NVIDIA Blackwell architecture at SIGGRAPH 2025. The lineup spans a compact system, a high-performance workstation and an enterprise-class server, each optimized and integrated with the NVIDIA Artificial Intelligence software stack. The company says the aim is to let designers and researchers manipulate and engage with digital content in real time, covering use cases that include video processing, Artificial Intelligence research and visual computing workflows.
The centerpiece of the announcement is the GIGABYTE XL44-SX2 server, positioned as a modular, future-proof platform for demanding visual and computational workloads. It pairs a scalable chassis design with the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, which Giga Computing highlights as targeted at enterprise workloads such as scientific computing, graphics rendering and video applications. The server supports up to eight GPUs and includes an integrated NVIDIA MGX PCIe switch board together with NVIDIA ConnectX-8 SuperNIC connectivity. Each GPU can be paired with 800G networking, a configuration intended to improve throughput and enable better scale-out across racks and clusters.
Beyond the server, Giga Computing will demonstrate systems aimed at different points on the performance spectrum, from space-conscious mini PCs to workstations built for heavier interactive and creative tasks. The company also notes its background in generative Artificial Intelligence servers and advanced cooling technologies, which inform the hardware configurations on display. By integrating GIGABYTE hardware with the NVIDIA software stack the vendor positions the systems for turnkey deployment in studios, research labs and enterprise environments. Attendees at SIGGRAPH 2025 will be able to see the hardware in person at the GIGABYTE booth and evaluate how the combination of Blackwell GPUs, MGX switching and high-speed networking can accelerate visualization and compute-heavy pipelines.
