Fujitsu is evaluating Rapidus as a potential manufacturer for its 1.4nm-class CPUs designed to power Japan’s next flagship supercomputer, marking a strategic shift toward domestic chip production, Nikkei XTECH reported.
The processors, dubbed FUJITSU-MONAKA-X, would support the Fugaku NEXT supercomputer slated to go online around 2030. The system aims to deliver performance gains of up to 100 times over the current Fugaku machine, developed in collaboration with NVIDIA and research institute RIKEN, according to the report.
While Fujitsu’s earlier MONAKA chip relies on TSMC’s 2nm process for a 2027 launch, the MONAKA-X represents a more ambitious technological leap. Key specifications are expected to be finalized by March 2026, with both TSMC and Rapidus under consideration for production.
The decision carries implications beyond a single project. Fujitsu is reportedly planning an investment in Rapidus, positioning the fledgling foundry as a candidate for future products as Japan seeks greater semiconductor independence.
Yet Rapidus faces formidable challenges. The company only began 2nm prototyping at its Hokkaido facility in July and targets mass production in 2027—the same timeline as TSMC, which already has established customers. Intel has moved faster, having unveiled its 18A-based Panther Lake chip earlier this month, while TSMC plans to commence 2nm production by year-end.