Tuesday, November 17, 2026
Privacy-First Edition
Back to NNN
Business

MiTAC Computing Scales Agentic AI with Full-Stack Infrastructure at SC26

Via CNW Group

MiTAC Computing Scales Agentic AI with Full-Stack Infrastructure at SC26 CNW Group Mon, November 16, 2026 at 10:15 PM EST 6 min read 3706.TW -0.16% AMD -3.31% INTC -7.73% 3706 NVDA -3.21% Explore stocks on Coinbase Trading disclosure Trading disclosure The above button links to Coinbase. Yahoo Finance is not a broker-dealer or investment adviser and does not offer securities or cryptocurrencies for sale or facilitate trading. Coinbase pays us for certain activity generated through this link. Prices displayed are informational.

CHICAGO, Nov. 16, 2026 /PRNewswire/ -- MiTAC Computing Technology Corporation, a subsidiary of MiTAC Holdings Corporation (TWSE: 3706) and a global leader in high-performance server solutions, today unveiled its comprehensive AI infrastructure ecosystem at SC26. At Booth 4615, themed "Scaling Agentic AI. Built on Full-Stack Infrastructure," MiTAC demonstrates how full-stack innovation overcomes the compute, thermal, and data bottlenecks of enterprise AI deployment. As Agentic AI drives an exponential surge in inference workloads, scaling infrastructure demands more than just raw GPU power. Through strategic collaborations with AMD, Canonical, DDN, Intel, NVIDIA, and Rafay, MiTAC delivers scalable AI factory architectures combining high-density liquid cooling, disaggregated inference, and end-to-end orchestration.

"Scaling Agentic AI requires cohesive infrastructure, not fragmented systems," said Rick Hwang, President of MiTAC Computing Technology. "MiTAC provides the foundational full-stack solution to make sustainable AI factories a reality."

Pioneering the Disaggregated AI Inference Engine

MiTAC and Skymizer are collaborating on an inference architecture that assigns compute-intensive prefill to GPUs and memory-intensive decode to Skymizer's HTX301 accelerator cards. The approach is designed to let enterprises provision each stage for its workload while keeping model inference on premises.

The proposed configuration combines two server platforms:

Prefill Stage: The MiTAC G8825Z5 (8U), featuring AMD EPYC™ 9575F or 9755 processors and AMD Instinct™ MI350X GPUs, is proposed for prefill processing and transfer of the resulting Key-Value (KV) cache to the decode server over high-speed networking.

Decode Stage: The air-cooled MiTAC G4520G6 (4U), equipped with Intel® Xeon® 6 processors, is proposed with eight Skymizer HTX301-4S accelerator cards, each with 256 GB of memory, providing 2 TB of aggregate accelerator memory across eight PCIe slots.

Together, the collaborative MiTAC G4520G6 decode server is designed to run massive models—such as the 2.8-trillion-parameter Kimi K3 large language model—on-premises to drive complex, deep-reasoning agentic AI workloads. Visit the landing page to learn more.

Orchestrating the Full-Stack Software Ecosystem

A scalable AI factory requires a cohesive software ecosystem. For foundational security, Canonical Ubuntu provides the certified operating system layer, ensuring seamless deployment and enterprise-grade reliability. Integrating the DDN Infinia solution ensures monumental data streams are processed with ultra-low I/O latency, preventing accelerator data starvation. Building on this foundation, MiTAC collaborates with Rafay to empower organizations with dynamic GPU-as-a-Service (GPUaaS) orchestration, efficiently allocating GPU resources across distributed workloads to eliminate idling and maximize ROI. Binding this ecosystem together is MiTAC's proprietary MiCoreView™ software, a comprehensive POD management solution granting deep telemetry and unified visibility across the data center.

Showcasing Diversified AI Infrastructure Through Rack-Scale Innovations

Building on AMD Chair and CEO Dr. Lisa Su's vision announced at OCP Global Summit last month, MiTAC is collaborating with AMD to develop the Helion Agentic AI Rack. Purpose-built for next-generation rack-scale agentic architectures, the upcoming ORv3 44OU Helion solution will scale up to 32,768 AMD Zen 6 CPU cores within an open-source stack—empowering hyperscale and enterprise environments to intelligently route complex workloads and maximize agents-per-watt efficiency.

Live on the SC26 show floor, MiTAC is also showcasing three rack-scale innovations built for modern workloads:

52U High-Density AI Liquid-Cooled Rack: Integrating 12 MiTAC G4826Z5 servers with 96 AMD Instinct™ MI355X GPUs, this solution achieves 50% higher GPU density than conventional deployments.

Advanced Liquid Cooling Rack: Partnering with ZutaCore, MiTAC exhibits a PoC integrating the G4826Z5 with Waterless Two-Phase Direct Liquid Cooling (DLC). Vaporizing dielectric fluid directly at the chip supports thermal design powers exceeding 4,000W and a heat flux over 250W/cm², enabling an annualized pPUE of 1.03.

High-Performance Storage Rack: Designed to feed Agentic AI without I/O bottlenecks, MiTAC's storage rack—powered by the DDN Infinia intelligent AI data platform utilizing seven GC68C-B8056 1U storage servers—delivers ultra-high-throughput, low-latency performance for continuous data streams.

Expanding the Foundation with Comprehensive Server Platforms

Beyond rack-scale architecture, MiTAC offers comprehensive server platforms for customized enterprise deployments:

G4827Z6: A 4U modular enterprise AI server leveraging NVIDIA MGX architecture and AMD EPYC™ 9006 SP7 processors. It pairs optimally with NVIDIA RTX PRO 4500 and 6000 Blackwell Server Edition GPUs for diverse inference workloads.

G2610Z6: A compact 2U edge GPU server with AMD EPYC™ 9006 SP8 processors. It seamlessly integrates with NVIDIA RTX PRO 4500 and 6000 Blackwell Server Edition GPUs, bringing robust AI capabilities to edge locations.

TN85-B8261: A dual-socket platform supporting the AMD Instinct™ MI350P PCIe Card, delivering highly efficient AI acceleration for specialized applications.

R1917GC: Serving as the orchestrating brain of edge data centers, this 1U management server utilizes NVIDIA Vera and NVIDIA Grace CPUs to coordinate complex clusters without internal GPUs.

R1610Z6: Featuring AMD EPYC™ 9006 SP8 processors in a dense 1U footprint, it delivers uncompromising single-socket performance for cloud-native workloads.

M2810Z6: A 2U2N multi-node server with AMD EPYC™ 9006 SP7 processors, optimizing compute density for software-defined hyper-converged infrastructure (HCI).

C2611Z6: A liquid-cooled multi-node server designed for OCP ORv3 racks and powered by AMD EPYC™ 9006 SP8 processors. It integrates four compute nodes in a compact 2OU form factor, each supporting 16 DDR5-8000 RDIMM slots and OCP 3.0 expansion for high-density, liquid-cooled infrastructure.

Attendees are welcome to visit MiTAC Computing at Booth 4615 from Nov. 17-19 at McCormick Place, Chicago.

About MiTAC Computing Technology Corporation

MiTAC Computing Technology Corp., a subsidiary of MiTAC Holdings, delivers comprehensive, energy-efficient server solutions backed by industry expertise dating back to the 1990s. Specializing in AI, HPC, cloud, and edge computing, MiTAC Computing employs rigorous methodologies to ensure uncompromising quality—across barebones, systems, racks, and cluster levels—fully achieving performance and integration. This commitment to quality at every level sets MiTAC Computing apart in the industry.

With a worldwide presence and end-to-end capabilities—from R&D and manufacturing to global support—MiTAC Computing provides agile, customized platforms for hyperscale data centers, HPC, and AI applications, ensuring optimal performance and scalability to meet unique business needs. By leveraging the latest advancements in AI and liquid cooling, and unifying the MiTAC brand with Intel DSG and TYAN server products, MiTAC Computing stands out for its innovative, efficient, and reliable server technology as well as its hardware and software integrated solutions—empowering businesses to meet future challenges.

MiTAC Computing Technology Corporation website: https://www.mitaccomputing.com/

AMD, the AMD logo, EPYC™, AMD Instinct™, and combinations thereof are trademarks of Advanced Micro Devices, Inc.

View original content to download multimedia:https://www.prnewswire.com/news-releases/mitac-computing-scales-agentic-ai-with-full-stack-infrastructure-at-sc26-302941327.html

View original content to download multimedia: http://www.newswire.ca/en/releases/archive/November2026/16/c5448.html

Read original at Yahoo Finance News

The Perspectives

0 verified voices · Three viewpoints · Real discourse

Left
0
Be the first to share a left perspective
Center
0
Be the first to share a center perspective
Right
0
Be the first to share a right perspective

Related Stories