Journalist and talking heads on TV are told exactly what vocabulary to use when covering a specific topic. It’s more directed at the interest of those in power and those with wealth than it is to fit an algorithm.
Don't forget heat. You can't just drain the local water supply to cool all your systems in space, you need to actually radiate all those kilowatts of power after your chips convert them into heat.
A lot of it hasn't actually been made, though. The AI companies have put in orders for future production. That future capacity can be redirected with a wave of a pen.
Continuing to operate nuclear plants that we've already paid for and are on the hook for decommissioning costs is perfectly sensible. Building new ones is what isn't.
There's lots of newer designs we could use, but they're still not going to be economically viable. A new nuclear plant isn't about today, it needs to be viable over its expected lifespan: around 2036 to 2076.
The fancy new designs aren't up against today's batteries. They're up against 2040's batteries, and they can't compete. For the price of a new nuclear plant, we'll be able to buy those massive battery farms and have money left over. Not today maybe, but a new reactor isn't going to start feeding power into the grid for ten years or so, so it'll need to compete with 2036's battery prices on day one and it's only going to get worse from there.
The main issues with nuclear power are that it's more expensive than renewables with battery storage, and that it's very slow to build which makes it the go-to option for coal and gas supporters looking to delay the energy transition for a few more years and use up all the funding.
There's probably a lot of people on 'stable' distros who are still running Wayland code from a couple of years ago and hitting bugs that have been fixed already.
Okay, but buying terrible, expensive vehicles from companies that fight tooth and nail to continue focusing on huge petrol SUVs isn't going to help.