Executive Summary
The accelerating progress of artificial intelligence (AI) and high-performance computing (HPC) is radically transforming enterprise IT and cloud infrastructure. As organizations demand ever-greater compute density, scalability, and energy efficiency, the need for an end-to-end, integrated hardware and software approach becomes mission-critical.
This white paper details how MiTAC Computing Technology, in partnership with AMD, delivers next-generation server platforms and data center solutions optimized for today’s most demanding AI, HPC, and data analytics workloads. We spotlight our leadership in rack-scale deployment, multi-node server design, and liquid cooling, all integrating the latest AMD EPYC™ CPUs and Instinct™ GPUs.
Introduction: Transforming the AI Data Center
AI and data-centric applications are shaping the modern digital economy. Real-time insights, generative AI, and deep learning require robust compute resources and flexible infrastructure. Enterprises and cloud providers need:
- CPU and GPU co-innovation for optimized workloads
- High-efficiency, high-density servers for scalability
- Openness—leveraging industry-standard networking architectures and the most popular software frameworks
- Advanced thermal and power management solutions
- Diversified supply chain to minimize vendor lock in
In partnership with AMD, MiTAC is uniquely equipped to address these requirements for global customers from cloud hyperscalers to enterprise IT environments.
AMD & MiTAC: Driving Innovation—Better Together
Integrated Platform Leadership
MiTAC’s server portfolio harnesses AMD’s leadership in CPUs and GPUs , networking, and open software:
- CPUs: AMD EPYC™ 9004/9005 Series—world-class performance and memory bandwidth for AI, HPC, and cloud
- GPUs: AMD Instinct™ MI300, MI325X, MI350 Series GPUs—extreme compute density for AI training and inference
- Networking: AMD Pensando™ Pollara NICs and DPUs which support Ultra Ethernet, UALink, and 400GbE-scale switching
- Software Ecosystem: AMD ROCm™ open stack which is completely up-streamed to support PyTorch, TensorFlow, JAX, and ONNX frameworks, for rapid AI innovation
Comprehensive Solutions for Demanding Data Centers
End-to-End Leadership - MiTAC and AMD together provide:
- Complete AI-ready server platforms, built for performance
- Flexible cooling—air and direct-to-chip liquid cooling (DLC)
- Scalable deployment options: rack-level to cluster-scale integration
- Next-generation networking with 400GbE capabilities
Product Highlights
AMD-Based GPU Servers: Liquid-Cooled and Air-Cooled
Liquid-Cooled MiTAC G4826Z5:
- 4U/19" form factor
- AMD EPYC 9004/9005 CPUs
- Up to 8 x AMD Instinct MI355X GPUs
Air-Cooled MiTAC G8825Z5:
- 8U/19" form factor
- AMD EPYC 9004/9005 CPUs
- Up to 8 x AMD Instinct MI300X/MI325X/MI350X GPUs
- Capable of 480TB storage and 400GbE fabric interconnects per rack
- Designed for both rack-level and cluster-level deployments; ideal for modern data centers’ AI and HPC workloads
Direct-to-Chip Liquid Cooling Solutions
- Complete rack-scale architecture: Up to 130kW per rack (64 x MI355X GPUs)
- In-row or in-rack CDU solutions
- Enables higher density deployments while maximizing energy efficiency
Multi node & OCP-Ready Platforms
TD76-B8058-D2/X4
2U with up to 2 or 4 EPYC nodes; PCIe Gen 5.0, DDR5 memory, and NVMe/E1.S for ultimate flexibility
TN85-B8261
2U/2S, dual AMD Turin sockets, supporting up to 24 NVMe drives and multiple GPUs per node
HG68-B8016-D5
6U/5N, optimized for AMD Ryzen, EPYC 4004/4005, and large-scale distributed computing needs
OCP ORV3 Platforms
fully compatible with open compute standards, delivered in both air and liquid-cooled configurations
Customer-Ready: Proven, Available, Evaluated
- Available now: AMD Instinct MI325X GPUs - Air-Cooled
- Remote POC available: AMD Instinct MI350X GPUs – DLC for customer benchmarking and validation families
- Flexible design: Custom rack, cluster, and management node configurations upon request
- Full-stack: Servers integrate AMD EPYC CPUs and optional Pensando AI NICs
Summary: Your Trusted AI Infrastructure Builder
MiTAC Computing delivers award-winning, production-ready platforms for AI and data center transformation:
- In close collaboration with AMD, our servers combine the latest CPU, GPU, networking, and cooling technologies
- Rack and cluster-level deployment, optimized for speed, flexibility, and energy efficiency
- Ongoing roadmap leadership—future-proofed for next-generation AMD processors and data center GPUs
For all your compute and AI server needs, contact MiTAC Computing Technology today.
TD76-B8058-D2/X4
2U with up to 2 or 4 EPYC nodes; PCIe Gen 5.0, DDR5 memory, and NVMe/E1.S for ultimate flexibility
TN85-B8261
2U/2S, dual AMD Turin sockets, supporting up to 24 NVMe drives and multiple GPUs per node
HG68-B8016-D5
6U/5N, optimized for AMD Ryzen, EPYC 4004/4005, and large-scale distributed computing needs
OCP ORV3 Platforms
fully compatible with open compute standards, delivered in both air and liquid-cooled configurations