Intel and AMD processors are the most common in home computers. These are manufactured in 5 and 7 nm like the Ryzen 7000/8000/9000, while other models for smartphones are usually around 3 and 4 nmFor example.
However, when it comes to supercomputers, data centers, advanced simulations and impressive artificial intelligence, the manufacturing process must be different.
The fewer the number of nanometers (nm), the greater the energy efficiency, speed, cores and density. Until now, the previous leader was Apple with its 3 nm for the A17 Pro from the iPhone 15 Pro and the M3 series for macbooks.
Fujitsu and Broadcom have achieved a milestone because the 2nm of its new chips is something that no other big brand has achievednot even NVIDIAor Samsung. Such innovation could change the technology industry forever.
Fujitsu-Monaka: the processor powering a new era of efficiency in supercomputers and AI
Fujitsu has shared its new great project on the official website, together with Broadcom and TSMC they are putting manufacturing of its first 2nm chips underway.
Mainly known as «FUJITSU-MONAKA»propose a new form of production that would be crucial to obtain efficiency benefits in various technological areas.
As mentioned, Apple had been focusing on 3nm and Samsung too, although with less manufacturing capacity, but it doesn’t all end there.
There is currently a race to reach 2 nm. For example, Apple A20 for the iPhone 18, Samsung Exynos 2600 for the Galaxy S26, AMD Zen 6 and Intel Nova Lake are expected to achieve it.
However, most have not been launched on the market, are in the testing stage or are focused on specific sectors such as smartphones and home PCs.
What the Fujitsu and Broadcom innovation proposes is a 144-core SoC with 3.5D packaging architecture (XDSiP) intended for Japan’s next flagship supercomputer FugakuNEXT directed by RIKEN.
On the other hand, it could be another blow to the current RAM memory crisis, since because it has to maintain performance in servers and supercomputers, it incorporates a memory subsystem with twelve DDR5 channels. Combined with PCIe 6.0 and CXL 3.0, it becomes an ideal resource for gigantic AI loads.
When will the Fujitsu-Monaka and in which sector it actually competes?
Of course, it is easy to compare it with other brands, but you have to keep in mind that there are many types of chips and several markets that they have to focus on.
The features of this incredible component offer reduced power consumption, increased signal density, It is faster and provides other long-term benefits.
This positions it as a competitor in the field of HPC (High Performance Computing) and high-level artificial intelligence. So, the main rivals it would have would be with AMD with Instinct, NVIDIA with Grace, Grace and Hopper or Intel with Xeon and Gaudi.
Is Fujitsu-Monaka the first in history with 2nm? It’s not the first, but yes the only one that is manufactured NOW and it’s not just advertised like most. Other brands are developing 2nm chips, but with other approaches.
Its design has been focused on the aforementioned client to finalize the FugakuNEXT and could be reaching data centers by 2027.