According to The Register, Fujitsu and Supermicro are teaming up to develop a new platform featuring Fujitsu’s upcoming high-performance, Arm-based MONAKA processor and advanced liquid cooling systems.

On October 3, Fujitsu and Supermicro announced a long-term partnership to develop and market a platform using the FUJITSU-MONAKA processor, which is designed for high performance and energy efficiency and set to launch in 2027. The companies will also collaborate on liquid-cooled systems for high-performance computing (HPC), generative AI, and eco-friendly data centers.

The Register highlights two key aspects of the alliance: Arm chips generally run cooler than competitors, requiring less thermal management innovation, and Fujitsu had largely exited this server market two years ago.

The Register reported that in February 2022, Fujitsu announced it would stop manufacturing and selling mainframe systems by 2030 and phase out Unix servers by the end of 2029. However, a year later, Fujitsu unveiled MONAKA, the successor to its A64FX chip, as a more energy-efficient, high-performance solution for HPC, AI, and data analytics workloads.

According to the companies, FUJITSU-MONAKA is based on the Arm architecture and employs cutting-edge 2-nanometer technology, with a 2027 release planned.

The partnership will combine Supermicro’s “Building Block” modular design, which allows customers to choose components optimized for their specific workloads, including cooling options like air-conditioned, free air cooled, or liquid cooling.

Fujitsu’s subsidiary, Fsas Technologies, will provide global generative AI solutions that integrate Supermicro’s GPU servers and support services for data centers and enterprises. Fsas was spun off last December to handle Fujitsu’s PC, server, and storage business, excluding mainframe and Unix systems.

The Register also mentioned that while liquid cooling for Arm systems is not common, it’s not unprecedented. For example, Fujitsu’s Fugaku supercomputer uses liquid cooling for its A64FX processors. This focus on liquid-cooled systems targets the growing demand for machines handling higher AI and HPC workloads.

(Photo credit: Supermicro)