Ask HN: Will quantum computing offset the power grid needs of AI?

Relevant link: https://www.nature.com/articles/s43588-023-00459-6

Currently most quibits require cryogenic cooling so without a breakthrough in room temperature superconductors etc. we could easily end up in a situation where quantum computing is used for specialized tasks where it is faster than conventional computing but the end result is more energy expenditure overall.

Edit: clarified the wording

It's hard to be sure, but Quantum Computing is useful only in a few specific problems, for example to calculate the Fourier Transform, because it's easy to parallelize and each part needs a very similar operations. With some clever trick, this can be used to factorize numbers. But it's not clear how to se it to solve arbitrary problems, so I guess a quantum AI is very difficult.