Gigabyte’s enterprise subsidiary, Giga Computing, has taken to the stage at SCA/HPC Asia 2026 to showcase its latest advancements in high-performance computing. The star of the show is undoubtedly the new XN24-VC0-LA61, a liquid-cooled server designed to bring the immense power of the NVIDIA Blackwell architecture to organisations that might not be ready for a full-scale rack deployment just yet.
The Gigabyte XN24-VC0-LA61
The XN24-VC0-LA61 is a 2U dual-processor server platform built on the NVIDIA MGX modular architecture. By utilising Direct Liquid Cooling (DLC), Gigabyte has managed to pack an incredible amount of hardware into a relatively small footprint. This modular approach allows for “Blackwell-class” computing power without the immediate need for a massive, full-rack infrastructure, making it a more accessible entry point for AI development.
Key Specifications and Features:
- Extreme Density: Powered by the NVIDIA GB200 NVL4 system.
- Processors: Two NVIDIA Grace CPUs (Arm-based).
- Graphics: Four NVIDIA Blackwell GPUs.
- Memory: 480 GB of LPDDR5X ECC CPU memory per Superchip and 186 GB of HBM3E GPU memory.
- Networking: Supports NVIDIA Quantum-X800 InfiniBand or Spectrum-X Ethernet (up to 800 Gb/s).
- Storage: Twelve PCIe Gen 5 NVMe drive bays.
- Power: 80 Plus Titanium redundant power supplies for maximum efficiency.
Tell Me More
While the liquid-cooled XN24 is the highlight, Gigabyte is also displaying a broader range of AI and HPC solutions. This includes the GIGAPOD, a massive rack-scale solution that integrates 32 GPU servers for “AI factories,” and the XL44-SX2-AAS1, an air-cooled modular server designed for enterprise training and visual computing using NVIDIA RTX PRO 6000 Blackwell cards.
For those looking at hyperscale data centres, Gigabyte also introduced the TO25-ZU4 and TO25-ZU5 compute nodes. These are compliant with Open Rack v3 standards and utilise the latest AMD EPYC 9005/9004 series processors, offering high scalability for cloud service providers.
What We Think
It is impressive to see how quickly Gigabyte is adapting the NVIDIA Blackwell architecture into various form factors. The XN24-VC0-LA61 is particularly interesting because it bridges the gap between standard servers and massive AI clusters. By using liquid cooling, they can keep these high-performance parts running efficiently without the noise and space requirements of traditional air-cooled setups. It’s a smart move for businesses looking to scale their AI capabilities without overhauling their entire server room.
Where to Buy
The XN24-VC0-LA61 and the broader GIGAPOD solutions are enterprise-grade products. Interested parties should contact Gigabyte’s enterprise sales team or authorised B2B distributors for regional pricing and availability.
