Fujitsu MONAKA Sales Start in November

Fujitsu MONAKA, the Japanese company’s new 2-nanometre Arm server processor, is scheduled to enter global sales in November 2026 alongside a made-in-Japan MONAKA Server. Fujitsu says the CPU will be sold both as a standalone component and inside its own server systems for AI, data-centre and high-performance-computing workloads.

What the Fujitsu MONAKA launch includes

Fujitsu’s distributed September 14 release says the processor reaches a maximum operating frequency of 3.8GHz and supports memory transfer speeds up to 8,800MT/s. The company is positioning the chip for AI inference and general data-centre computing rather than only for a successor to the Fugaku supercomputer.

The sales model matters as much as the specifications. Fujitsu says cloud operators, data-centre providers and server vendors will be able to buy the processor separately. Its own MONAKA Server will target customers in Japan and Europe, giving the company a route to sell both silicon and a finished infrastructure system.

The sovereignty pitch is a supply-chain pitch

Fujitsu describes the product as sovereign AI infrastructure because the processor is designed in Japan and the server is integrated, developed and manufactured there. That does not mean every component is domestic: independent reporting from Seoul Economic Daily says TSMC manufactures the 2nm chip under contract.

The distinction is useful for buyers. Sovereignty in this case is about design control, system integration, traceability and operating autonomy, not a claim that the complete semiconductor supply chain sits inside Japan. Fujitsu’s earlier February manufacturing announcement also linked the server programme to confidential computing and critical-infrastructure requirements.

Element	Company plan	Evidence boundary
CPU sales	Global from November 2026	Forward-looking launch schedule
Server availability	Japan and Europe from November	Company announcement
Process	2nm compute technology	Independent production benchmarks pending
Clock / memory	Up to 3.8GHz / 8,800MT/s	Vendor specifications

What buyers should watch next

The key test is whether November brings production systems and repeatable workload results, not only engineering samples. Tom’s Hardware’s August architecture reporting described a 144-core Arm design with a separate 5nm cache die, but noted that performance claims would need independent testing.

Procurement teams will also need pricing, software-compatibility matrices, delivery lead times and a clear support model. Fujitsu’s announcement establishes the route to market, but it does not yet provide enough public evidence to calculate total cost of ownership against established Arm or x86 platforms. Those operational details will decide whether MONAKA moves beyond specialist HPC and sovereign-computing buyers.

Fujitsu MONAKA enters a market where large cloud companies increasingly design their own Arm processors and GPU systems dominate AI training. Its opening is narrower: efficient inference, HPC and sovereign deployments where architecture control matters. NVIDIA’s CUDA-Q Logical launch similarly shows infrastructure vendors competing through orchestration layers, while the Equatys satellite venture illustrates how control of physical networks is becoming part of technology strategy.

Frequently asked questions

When will Fujitsu MONAKA go on sale?

Fujitsu says global standalone CPU sales and broad MONAKA Server availability will begin in November 2026.

Is Fujitsu MONAKA manufactured entirely in Japan?

Fujitsu says the server is integrated and manufactured in Japan, while independent reporting identifies TSMC as the contract chip manufacturer.

Are the performance claims independently verified?

Not yet. Clock, memory and efficiency figures should be treated as vendor specifications until production systems are benchmarked independently.

Sources: Fujitsu release distributed by PR Newswire/AAP; Fujitsu MONAKA product page; Seoul Economic Daily; Tom’s Hardware.