Fujitsu formally announced its MONAKA CPU, a 144-core Arm-based processor designed for sovereign AI infrastructure. The CPU utilizes 2nm 3D-stacked technology and is slated for availability beginning in November.
The new Fujitsu MONAKA CPU and server are designed to deliver "world-class AI inference performance." According to Fujitsu, the CPU features a maximum operating frequency of 3.8GHz and supports memory speeds up to 8800MT/s. The architecture combines a 2nm process for the core with a 5nm process for the cache and I/O sections, enhancing power efficiency and cost-effectiveness through the company’s ultra-low voltage operation technology.
Fujitsu states the MONAKA CPU provides hardware acceleration for matrix operations central to AI inference, using dedicated instructions. Combined with SVE2 vector operations and software optimization, it achieves twice the throughput for AI inference compared to other CPUs, enabling practical AI inference performance even in locations with power and cooling constraints.
The 1U air-cooled rackmount server will feature 144 ARMv9 cores running at 2.1GHz, or 2.9GHz with liquid cooling. The 2U rackmountable air-cooled server also operates with 144 cores at 2.1GHz. Both server configurations support twelve RDIMMs per socket and PCIe Gen6. Initial availability begins in November, with expanded availability in Europe planned for 2027.
Read the original coverage
💬 Comments
📜 Comment Policy