One of my favorites is Rockchip RK3588. 8nm process, quad-core Cortex-A76 + quad-core Cortex-A55 ARM Mali-G610 MC4 GPU. I have some SBC with it. I like it!The way I see it, there is a lot a middle ground between the 16nm process used for the Raspberry Pi 5 and a leading-edge 2nm node. For example, the quad-core Amlogic S905X5M is on 6nm and runs amazingly cool. However, it's not very fast.The other potential advantage of smaller silicon is cutting more chips from the one silicon sheet. I do not know how that compares to the increased cost of tighter lithography.
Note that the RP1 GPIO controller on the Pi 5 is fabricated using a 40nm process node, which from what I understand is important for driving signals over chunky pins in physical computing projects. While moving GPIO off the SOC wasn't ideal from a latency point of view, this definitely opens the door to using smaller nodes that are faster and lower power for the processor.
To answer the question directly, a 2nm process node would allow making a much faster ARM CPU with lower power dissipation--especially with the newer Cortex-X processor designs.
What is possible and what is practical for a Pi are, of course, two different things. At the same time, I have no idea what will happen in the future. It's possible the chatbots will evolve into humanoid robots. It's possible there'll be a devastating war. Maybe these two outcomes are actually the same.
Statistics: Posted by Olydnad_SWE — Mon Feb 09, 2026 8:55 am