Built for large-scale AI infrastructure with liquid- and air-cooled platforms powered by AMD Instinct MI355X GPUs
AMAX announced support for the AMD Instinct™ MI350 Series platform across its next-generation server and rack-scale solutions. Designed to meet the compute demands of today’s AI workloads, the new AMAX offerings deliver increased performance, memory bandwidth, and deployment flexibility for training, inference, and data-intensive computing.
AMAX’s engineering-led approach and support for both server and rack-scale liquid-cooled solutions make them a strong partner for customers looking to deploy AMD-based AI infrastructure at scale.”
— Ted Marena, Director of AI at AMD
LiquidMax® RackScale 128: Liquid-cooled 51OU rack supporting up to 128x MI355X or MI325X GPUs, featuring cold plate cooling, CDU support, and centralized 48V power for scalable rack-level AI deployments.
AceleMax® AXG-728AM2: 7U AMD EPYC™ 9005 server supporting 8x MI355X OAM GPUs, built for high-density AI training and inference with up to 3TB DDR5 memory and 11 PCIe Gen5 slots.
Also Read: CIO Influence Interview with Dipto Chakravarty, Chief Product and Technology Officer at Black Duck
AMAX will offer full platform support across the MI350 Series, with solutions optimized for both air and liquid-cooled environments.
Based on AMD’s CDNA™ 4 architecture, the MI350 Series delivers 288GB of HBM3e memory per GPU, expanded FP6 and FP4 support, and up to 1.8x FP16/FP8 throughput over the previous generation. These improvements help organizations accelerate AI workloads while improving compute efficiency and infrastructure readiness.
“AI infrastructure is under pressure to deliver more performance without expanding footprint,” said Rene Meyer, Chief Technology Officer of AMAX. “We engineered our MI355X solutions to meet that demand directly, combining compute density with serviceable thermal and power design that aligns with how our customers build for the future.”
Also Read: About IoT Security: Challenges and Tips for a Hyperconnected World
“With the AMD Instinct MI355X, we’re delivering the performance and memory capacity needed for next-generation AI models,” said Ted Marena, Director of AI at AMD. “AMAX’s engineering-led approach and support for both server and rack-scale liquid-cooled solutions make them a strong partner for customers looking to deploy AMD-based AI infrastructure at scale.”
All AMAX platforms featuring AMD Instinct MI350 Series GPUs are designed for rack-scale deployment and supported with services including performance validation, thermal optimization, and remote testing.
[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]
