Fujitsu Plans Global MONAKA AI Chip Launch With TSMC
Fujitsu is moving deeper into the data-center computing market with MONAKA, an internally designed processor aimed at artificial intelligence workloads. The chip will mark the Japanese technology group’s first commercial sale of an advanced AI processor to outside customers. Production by Taiwan Semiconductor Manufacturing Co. using its 2-nanometer process would pair Fujitsu’s chip-design expertise with one of the industry’s most advanced manufacturing technologies.
Fujitsu plans to begin exporting MONAKA to global markets in 2027 and has secured U.S. server maker Super Micro Computer as a customer, according to reports. The company also intends to sell AI servers equipped with the processor, expanding beyond standalone chip shipments into complete systems. Fujitsu is targeting cumulative MONAKA-related revenue of 250 billion yen by fiscal 2030 as it seeks a larger share of AI infrastructure spending.
All Coverage
2 original reportsThe Backstory
The history behind this eventFujitsu Unveils 144-Core Monaka Chip for AI Workloads
Fujitsu’s Monaka is a next-generation server processor designed for artificial intelligence inference and high-performance computing, where operators are seeking more processing capacity and better energy efficiency. Built on the Arm architecture, the project highlights Fujitsu’s push to develop proprietary silicon for data centers and strengthen Japan’s position in advanced computing infrastructure.
Fujitsu disclosed new Monaka details at Hot Chips 2026, including a 144-core design and compute cores manufactured with Taiwan Semiconductor Manufacturing Co.’s 2-nanometer process. The processor also combines 5-nanometer technology with a 3D multi-die packaging design. Fujitsu is targeting AI and HPC server workloads and expects volume shipments to begin in 2027.
Fujitsu Taps Rapidus 1.4-Nanometer Process for AI Inference NPU
Fujitsu is developing its in-house Monaka CPU for the AI infrastructure and supercomputing markets. It is now also adopting an advanced process from Japanese foundry Rapidus to develop an inference NPU, aiming to reduce reliance on high-end chips from overseas and strengthen Japan’s semiconductor supply-chain autonomy.
The latest plan calls for the AI inference NPU, built using Rapidus’ 1.4-nanometer process, to be integrated with the Monaka CPU in a single package, with supply expected to begin in 2029. The project has received funding from Japan’s New Energy and Industrial Technology Development Organization, or NEDO, but the amount of the subsidy and the exact mass-production date have not been disclosed.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →