Can't speak for Australia, but at least in the US, a big part of the reason rooftop solar ends up looking like such a savings is net electric metering, where power companies buy surplus power from solar panel owners at the same rates that they sell those users power at night. The tricky thing is that the price users pay for power doesn't just include generation costs, but also transit, grid maintenance, etc., which rooftop solar owners benefit from even though they no longer pay for it -- these costs end up being borne by other rate-payers, who effectively cross-subsidize rooftop owners' freeloading. In the early days of solar, there was a good public policy argument to be made here: solar power is a net societal good (air quality, carbon footprint, creation of solar economies of scale that drive prices down for everyone, etc.), and power companies weren't going to do solar on their own.

Now, though (again, at least in the US), utilities are making massive investments in grid solar, and it's dramatically cheaper than rooftop solar. Like: a quarter the cost per kilowatt hour of generation capacity. So now I think it's sensible to ask: if we're going to ask ratepayers to subsidize solar generation, does it really make sense for us to spend their subsidy on a way of doing solar that's objectively much more expensive just because it saves a few specific users money? They could be getting 4x the bang for their buck if we spent that subsidy on grid-scale instead, and then the economics of rooftop solar mostly go to shit (as well they probably should -- it's needlessly expensive).