Fujitsu announced its plans for what will be its next top-of-the-range CPU, the Fujitsu Monaka. Designed for data centers, Fujitsu seeks to create a powerful yet efficient processor, so they will resort to Armv9 architectureand TSMC’s most advanced manufacturing process, the 2nm.
Obviously, this presents a small problem for Fujitsu, and that is that this processor is expected come to life by 2027. So he still has a lot of time ahead of him to replace his Fujitsu A64FX. This is its current top-of-the-range processor for high-performance computing and artificial intelligence. Of course, it is already obsolete for this task with 48 cores Arm alongside a 7nm manufacturing process from TSMC.
This is what we know about the Fujitsu Monaka CPU
It was this week when Fujitsu announced that its future lies in the Fujitsu Monaka CPU. As indicated, we will find a processor 150 cores based on Armv9-A architecture. These cores will be linked to a manufacturing process of 2nm from TSMC, along with 3D assembly technology using chiplets. The curious thing is that it actually indicates that the number of cores of each of these CPUs is “approximately“150 cores.
On the other hand, RAM memory support could not be missing DDR5. Not to mention interface support PCI Express 6.0 complying with the Compute Express Link 3.0 (CXL 3.0) specification. It’s also no surprise that thanks to Arm and the manufacturing process, this CPU is designed to be air cooled.
The Fujitsu Monaka CPU will be manufactured sometime in 2026, but it will not be until 2027 that it will be available for use. This is the same year in which it is expected that Rapidus Start manufacturing semiconductors with a 2nm manufacturing process in Japan. But of course, arriving and trying to manufacture a 150-core CPU, with a 3D desEsports Extrasand chiplets, would be a quite ambitious plan for its beginnings in this lithography.
Fujistsu’s goals with this new CPU for data centers and AI
According to the company, the main objective of the Fujitsu Monaka is to make compatible “high performance with energy savings“. We continue with creating a “high speed data processing infrastructure“, and all this while being reliable, secure and easy to use thanks to the Arm software ecosystem.
“The amount of data handled worldwide in 2030 is projected to reach ten times the amount of data handled in 2020, and explosive growth is expected to continue after that. Demand for large-scale data centers will grow along with increasing data volume, but it will be difficult to achieve carbon neutrality unless data energy consumption is curbed.