Exactly the "too cheap to meter" claims that were made about fission energy.
The problem is that though the marginal cost of the next watt-hour of electricity may be low you still have to pay the very real (and likely huge) costs of constructing and maintaining the reactors and generating plant. Investors will need to earn a return; the bonds payments will need to be made.
Fusion (particularly DT fusion) is grossly overvalued as an energy source. It's been an element of the consensus futurology (which is why you see it as a trope in SF stories), but this placement is not deserved by its actual qualities.
Renewables make much more sense. At the experience curve for utility scale PV from 2010 to 2022, assuming that experience rate of 33.1% decline per doubling of cumulative deployment would continue, if the world were powered by PV the levelized cost of delivered power would fall to $0.003/kWh.
Not forever, but they're enough to save us from specifically the energy part of the problem, for the next 45 years even at PV growth rates, or for 390 years if it reduces to only 3% annual compound growth.