Laptops use lpddr, which uses much less power than a desktop dimm. That said they’re also saying in standby which represents and massive power reduction as clocks are dropped to minimum and the OS can shutdown unused banks.
So available battery/desktop dimm power usage is not the right maths to be using here.
I’m not saying they aren’t exaggerating or anything, but in general marketing information is always technically correct, even if you would not get that perf in real life so outright falsehood should indicate your base assumptions can be wrong
You can get a million hours of playback with enough battery. Thin and light laptops with typical sub-50w/hr batteries are going to have a very hard time going over 20-25 hours.
I'm also super-skeptical of AMD's claim because their chipsets have a very long history of sucking a lot of power.
He's talking about the chipset/motherboard/battery management logic/etc, which isn't in the CPU. Even then, you need the OS to properly utilize the logic, and Linux so often has worse power management than windows for laptops.
This really is a case where the CPU vendor (AMD) needs to get involved with the hardware partners to get good drivers made FOR LINUX to showcase their hardware, and not just windows.
The real reason devs love having Apple laptops isn't OSX. It isn't Apple the brand. It is somewhat the quality of hardware... but the real reason is that it is IT-supported Linux.
Looking ahead to the seemingly inevitable, AMD and Intel when pushing ARM (or RISC-V?) CPUs, will need to formally adopt good linux support for the hardware.
An instruction set changeover cannot succeed through non-technical consumers in Windows. It won't in games. The only beachhead for AMD/Intel in non-x86 will be a good Linux laptop. Otherwise Apple will have the advantage.
It seems pretty reasonable considering my 4-year old laptop with a "power hungry" 4K screen uses 4 watts in very light usage. For a start, they are probably using LPDDR5, just like phones.
So available battery/desktop dimm power usage is not the right maths to be using here.
I’m not saying they aren’t exaggerating or anything, but in general marketing information is always technically correct, even if you would not get that perf in real life so outright falsehood should indicate your base assumptions can be wrong