The big providers are losing on average tens of billions a year on these services, so yes prices must go up. Even Moore’s won’t help in the medium-term due to shortages and difficulty/reluctance to vastly increase capacity.
The frontier labs have very high prices for inference. The prices are actually going down, not up.
Others running open models in the cloud is a nice third alternative, but not a solution for those that need frontier models, which I believe continue to need training as well as inference. These are currently heavily subsidized and hardware constrained several years out.
And this sub-discussion is more about whether there will continue to be some cloud model that's both more capable and cheaper than local, not so much about what kind of cloud model it is.