Some stuff like the buildings and power supplies will probably remain good. But year, probably new chips in a short while.
We have been terrified to whisper the words "nuclear power" for decades now, but the AI boom is likely to put enough demand on the power grid that it forces us to face this reality and make appropriate buildouts.
Even if the AI Boom crashes, these power plants will have positive impacts on the country for decades, likely centuries to come. Keeping bountiful power available and likely low-cost.
-Elliot Aronson, "The Rationalizing Animal"
Or to be less philisophical: the people with the money and power to say what's important are rarely the ones thinking long term, nor in an audience of other powerful, rich people thinking long term. US congress' median age is over 60: most aren't thinking about how to keep the Earth alive in 20-30 years. They won't be around to suffer the consequences.
AI on the other hand promises immeidate gain(lot of expensive job automation = cost cutting) as well as future return on investment. Like someone said, IT has returns realized quickly in few years, so that is more lucrative than /reducing pollution/.
Also, reducing pollution requires money spent(cost) and everyone is afraid of the cost if it does not promise at least x2-5 minimum return gain immediately(or in few years).
Whereas exploiting a new technology like AI for potential profit is like a massive hit of sugar/caffeine/drug in that we feel/act on ASAP.
It makes more sense when you understand "training AI models" as "greedily pumping up a bubble."
Someone really important or really rich needs to build that demand so we get something for the wrong reasons but for potential good intent. FWIW, I'm not really optimistic that the bubble lasts long enough to even get these plants off the planning stage, though.
I don't really expect the current LLM bubble to last long enough to stand up new nuclear plants, though I don't expect that to actually stop the new energy projects unless the bubble popping has a massive economic impact.
Power plants are slow moving projects. Even if LLMs as they are don't live up to expectations they have seemed to open the door for the idea that amazing things are coming and we need to make sure the energy supply is ready for it.
I can see it going either way. It really depends on how the economy moves this decade, and I feel we're very much at an inflection point as of now. I won't even pretend to guess how 2025-6 will go at this point.