HOME News Center Back
Press Release
Nov. 18, 2026

Giga Computing Puts Rack-Scale Infrastructure and Agentic AI Center Stage at Supercomputing 2026

Share:
Banner_1030_SC

 

November 17, 2026 - Giga Computing, a recognized leader in AI hardware innovation and advanced cooling technologies, today announced its participation in Supercomputing 2026 (SC26) in Chicago. Under the theme “HPC Unites,” SC26 brings the global high-performance computing community to McCormick Place from November 16-19, where researchers, industry experts, and government organizations share the technologies that drive scientific and research progress. At GIGABYTE Booth #2229, Giga Computing shows how unified compute, liquid cooling, and partner ecosystems come together, with a full portfolio that runs from compact, personal AI supercomputers to rack-scale AI infrastructure.

⁠

Throughout SC26, the GIGABYTE booth will host a series of technical sessions led by Giga Computing and its industry partners, including AMD, Intel, NVIDIA, Phison, Schneider Electric, and WEKA. Covering the full AI deployment lifecycle, from AI Factory infrastructure planning to software in the AI stack and GPU allocation, the sessions offer attendees practical, end-to-end guidance on building and operating AI at scale. The full session schedule is available at https://gogct.link/GCTSC26.

⁠

GIGABYTE GIGAPOD: expanding rack-scale solutions

With the rise of hyperscalers and AI cloud providers, the data center has shifted focus to massive rack-scale deployments. Two years after launching GIGABYTE GIGAPOD, Giga Computing continues to expand its POD portfolio to offer more integrated solutions for Agentic AI and simplified management of resources. Shown at the GIGABYTE booth is a management rack solution that combines compute nodes, a simplified liquid & power management (G-rex) solution, switches, and servers to more efficiently feed data to GPUs.

  • GIGABYTE G-REX: simplifies management of CDUs, PDUs, RDHx, and leak detection into a simple and friendly interface.
  • UfiSpace provides switches for east-west data fabrics and AI training clusters, as well as switches for Out-of-Band management.
  • GPU to Storage Solution: Partners Vast and WEKA, provide ultra-fast, centralized storage for AI and HPC clusters with GIGABYTE servers.

⁠

High-Density, Liquid-Cooled CPU Compute Rack

New rack-scale blade solutions are at the GIGABYTE booth. A new 6U, liquid-cooled chassis integrates with CDU, PDU, manifold infrastructure, with cold plates on the CPU, memory, NICs, that capture nearly 90% of system heat. The GIGABYTE B685-D80-LS1 is built for AMD EPYC™ 9006 Server CPUs, delivering up to 5,120 CPU cores in a single server. Alongside it, the GIGABYTE G684-A80-LAS1 supports Intel® Xeon® 6 CPUs in the same blade server form factor. These servers integrate with cooling provided by an in-rack GIGABYTE CDU and power distribution from Schneider Electric, giving customers a simplified and validated process to deployment.

⁠

Newly introduced is the GIGABYTE H295-D80-CS1. This 2U, four-node closed-loop liquid cooling server supports two EPYC 9006-based CPUs per node for up to 2,048 CPU cores per server. By integrating a closed-loop system, the server can be more easily integrated into traditional data centers that do not rely on a connection to facility water, providing liquid-cooling efficiency without the infrastructure.

⁠

Infrastructure Projects

Visitors can discuss containerized, modular deployments and AI factories at the booth with the Giga Computing team. New reference designs have made it simpler to deploy clusters at scale. GIGABYTE is highlighting GAIFA (GIGABYTE AI Factory Accelerator) and GADU (GIGABYTE Accelerated Deployment Unit) as examples of quick to deploy AI infrastructure. Visit to learn more.

⁠

AMD-based Solutions:

  • G295-D21-AP1: GPU server using EPYC 9006 CPUs and AMD Instinct™ MI350P GPUs. This scale-up server is ready for next gen, PCIe Gen6, GPUs.
  • G4L3-ZX1-LAT4: Dual EPYC CPUs pair with AMD Instinct™ MI355X GPUs for massive LLM models and double-precision scientific simulations.
  • W793-ZU0-LA71: (Live demo) An AMD EPYC & Radeon™ PRO-based workstation shows how AI can be used for crowd monitoring and safety, while maintaining data sovereignty.

⁠

Intel-based Solutions:

  • G494-AB0: A prototype of an upcoming GPU server for Intel Xeon processors (Diamond Rapids) that supports 16-channel memory and PCIe Gen6 GPUs.
  • B343-X40-AAS1: Hosting/cloud server with ten blades in a 3U chassis, each having two nodes using the Intel Xeon 6300-Series processor.
  • R164-SG6-AAJ1: Supporting sixteen E3.S drives and additional PCIe slots in a 1U server for Intel® Xeon® 6700/6500-Series Processors with 136 PCIe lanes.

⁠

NVIDIA-based Solutions:

  • G3L4-SD5-LA08: First GIGABYTE server to launch in support of the NVIDIA HGX Rubin NVL8 platform. Featuring a unique 3U design, adding U.2 drives for KV caching.
  • AI TOP ATOM: (Live demo) mini AI supercomputer powered by NVIDIA GB10 Grace Blackwell Superchip and NVIDIA AI software. Supports up to 200B parameter models and achieves up to 1 petaFLOP of FP4 AI performance.
  • W775-V10: (Live demo) Deskside supercomputer accelerated by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip. Supports up to 1 trillion parameter models and achieves 20 petaFLOPS of FP4 computing power.
  • NVIDIA Vera Rubin NVL72: NVIDIA Vera Rubin NVL72 supercharges the era of agentic AI - delivering inference with 10x more throughput per watt, 1/10th the token cost and training with one-fourth the GPUs.
  • XV25-VC0-AJ1: Based on NVIDIA MGX architecture, the NVIDIA Vera CPU platform with support for four NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs.
  • Additional NVIDIA rack-scale solutions supported in future GIGABYTE solutions: NVIDIA Groq 3 LPX, NVIDIA Vera CPU, NVIDIA STX, and NVIDIA Rubin NVL4.

⁠

True to the “HPC Unites” theme, Giga Computing brings every layer of the data center together in one deployable solution. From deskside AI systems to dense CPU racks and AI factories built with GAIFA and GADU, Giga Computing delivers validated hardware, liquid cooling, and an open partner ecosystem. Visit Booth #2229 for live demos and expert guidance.

 

language

Under the General Data Protection Regulation (GDPR) enforced by the European Union, we are committed to safeguarding your personal data and providing you with control over its use.

Manage Cookies

Privacy preferences

Under the General Data Protection Regulation (GDPR) enforced by the European Union, we are committed to safeguarding your personal data and providing you with control over its use.

Privacy Policy

Manage preferences

Necessary cookie

Always on

By clicking "Accept All," you consent to our use of cookies to enhance your experience on this website, assist with performance analysis, and deliver relevant marketing content. You can manage your cookie preferences below. Clicking "Confirm" indicates your agreement to the current settings.