SAN FRANCISCO, July 24 2026 — MiTAC Computing Technology Corp. has introduced a new portfolio of AI server platforms powered by AMD’s latest 6th Generation EPYC server processors, expanding its infrastructure offerings for artificial intelligence (AI), cloud computing, high-performance computing (HPC) and enterprise applications.
The launch, announced at the AMD Advancing AI event, includes new liquid-cooled and air-cooled server solutions designed to deliver higher performance, improved energy efficiency and greater scalability for modern data centres.
Among the new products are the MiTAC M2810Z6 and MiTAC M2610Z6, both built on AMD’s sixth-generation EPYC processors. The platforms feature support for the latest “Zen 6” architecture, PCIe Gen 6 connectivity and CXL 3.1 technology, enabling higher compute density, faster memory bandwidth and expanded input/output capabilities for AI-driven workloads.
MiTAC President Rick Hwang said the latest platforms are designed to help organisations deploy AI-ready infrastructure while supporting existing enterprise applications.
The company also showcased the TN85-B8261, a dual-socket AI server supporting AMD Instinct MI350P graphics processors for AI training, inference, high-performance computing and data analytics. The system supports up to four PCIe 5.0 GPUs and up to 6TB of DDR5 memory.
In addition, MiTAC introduced a 52U high-density liquid-cooled AI rack capable of supporting up to 96 AMD Instinct MI355X GPUs. The company said the solution delivers up to 50 per cent higher GPU density per rack while reducing data centre footprint by as much as one-third compared with conventional AI deployments.
The company also presented its G8825Z5 AI server, which integrates Diamond Cooling technology to improve thermal efficiency and reduce energy consumption. According to MiTAC, the platform can deliver up to 50 per cent more AI processing performance per watt while lowering cooling requirements and operational costs.
To support AI development and validation, MiTAC has also introduced remote AI rack testing capabilities, allowing engineers and developers to evaluate workloads on shared infrastructure powered by AMD Instinct GPUs and EPYC processors.
The company said the expanded portfolio reflects growing demand for AI infrastructure capable of supporting increasingly complex enterprise, cloud and high-performance computing workloads while improving energy efficiency and lowering total cost of ownership.
–Minutesmy
