MiTAC Computing Technology Corp., a global provider of high-performance and energy-efficient server solutions and a subsidiary of MiTAC Holdings Corporation, has introduced its latest agentic artificial intelligence (AI) infrastructure at the AMD Advancing AI showcase.
Designed to cater to the changing needs of enterprises, the MiTAC booth will feature its latest lineup of its high-density liquid-cooled and air-cooled AI servers alongside their advanced multi-node platform based on the 6th generation AMD EPYC™ processor. This new set of enterprise products provides seamless workload migration between AI, cloud computing, HPC, and enterprise application solutions.
Driving AI Readiness and Data Center Efficiency
The introduction of MiTAC’s new server models, including the M2810Z6 and M2610Z6, underscores the company’s focus on delivering high-density, high-throughput data center infrastructure. Leveraging AMD’s “Zen 6” Venice architecture, PCIe Gen 6 I/O subsystem support, and CXL 3.1 standards, these hardware solutions provide the memory bandwidth and expansion capabilities needed to manage complex agentic AI and machine learning workloads.
“Agentic AI and next-generation enterprise workloads require a fast, scalable, and efficient infrastructure,” said Rick Hwang, president of MiTAC Computing Technology. “With the launch of our all-new platform powered by 6th-generation AMD EPYC server processors , MiTAC provides customers with the essential building blocks to deploy an AI-ready infrastructure while supporting the workloads that drive their businesses today.”
Core Product Highlights
-
High-Density Compute: MiTAC M2810Z6
- Architecture: 2U multi-node chassis with 2 single-socket nodes using 6th Gen AMD EPYC SP7 server processors.
- Power and Thermal: Designed for up to 600W Configurable Thermal Design Power (cTDP) support per node.
- Memory and I/O: 16 DDR5 DIMM slots for up to 4TB DDR5-8000 memory capacity.
- Storage and Networking: Hot swappable front NVMe drives (E3.S, E1.S, U.2) and OCP 3.0 TSFF extension support for cloud-based deployments.
-
Highly Flexible Enterprise Server: MiTAC M2610Z6
- Workload Applications: Virtualization, cloud storage, edge computing, and telecom services.
- Architecture: 2U 2-node platform with single-socket AMD EPYC™ 9006 SP8 processors (up to 400W cTDP).
- Memory and Storage: 16 DIMM slots with 4TB DDR5-6000 memory, hot-swappable NVMe storage, and OCP 3.0 connectivity.
“Enterprise AI has moved from the experimental stage to production, and agentic AI will place even greater demands on computing power, memory bandwidth, efficiency, and system-level orchestration,” said Dan McNamara, senior vice president and general manager of AMD’s Compute and Enterprise AI business. “With 6th generation AMD EPYC server processors , AMD is helping customers scale their enterprise AI with the performance and efficiency needed to support increasingly complex data pipelines, AI services, and intelligent applications in modern data centers.”
Also Read: IBM Unveils Autonomous IT Software and Next-Gen Power Systems to Drive Enterprise Resilience and Efficiency
Specialized GPU Acceleration & Advanced Cooling Innovations
Beyond general server architecture, MiTAC showcased high-density GPU platforms designed for large-scale AI training and inference:
- MiTAC TN85-B8261: A 2U Dual Socket AI Server with Dual AMD EPYC™ 9005/9004 Series Processors, 6TB DDR5 Memory, and 4x Double-Width PCIe 5.0 slots. Designed to accommodate AMD Instinct™ MI350P GPUs and AMD Pensando™ Pollara 400 AI NICs, it comes with 8 x hot-swap U.2 NVMe storage bays and Titanium redundant PSUs.
- 52U High-Density Liquid-Cooled AI Rack: A rack-scale system featuring 12 MiTAC G4826Z5 AI Servers with up to 96 AMD Instinct™ MI355X GPUs. It offers a 50% boost in GPU density per rack when compared to standard systems, reducing physical footprint by up to 33%.
- MiTAC G8825Z5 with Diamond Cooling®: Using Akash Systems’ unique thermal management capabilities, the server runs AMD Instinct™ MI350X GPUs and provides 50% more AI tokens per watt at operating temperature while avoiding any thermal throttling at inlet temperatures of above 95°F.
Collaborative Testing & Validation Facilities
In an effort to facilitate the deployment process for customers, MiTAC has deployed three separate AI test bed clusters with AMD Instinct™ MI350 GPUs and AMD EPYC™ processors. Ninety-six engineers and software developers can access this common system remotely to conduct benchmarking, validate models, and develop AI models collaboratively.
