I skimmed through this on another forum. It seems widely misunderstood.
It applies to high volume system manufacturers. Small builders (I think I saw 50 units mentioned) and self builders are not affected in any way.
It aims to reduce power consumption of PCs when they are idle or otherwise not in use. This includes on but not doing anything, sleep mode, and "off" modes. Even when "off" you usually still have various standby power circuits active, unless you pull the plug. It does not do anything about in-use power! So you can still run your 300W GPUs with 200W CPUs as long as you like. The power consumption is weighted, and practically speaking it was 35% idle, 65% sleep/off. So that is roughly comparable to 8 hours a day every day, or 12 hours for 5 days.
It recognises that more performant PCs will use more power than less performant PCs, so a high power workstation will have different limits than a thin client. This is where it gets really messy, as they give points for different classes, ports and some hardware specs. I think they had to do this, as opposed to a more sensible efficiency metric for notable components, since the latter would be impractical to enforce at a state level. It would have to be applied nationally to work.
At a practical level, this might start to kill off RGB from prebuilts, since that does not contribute to performance and will eat into the power allowance. I don't know if they can get away with it by shipping it RGB fitted but disabled. It is up to the user to enable it and burn that power. AIOs might be another victim to this.