You need fewer mounting systems if the production per panel is higher. It's one of the major things you can do to reduce costs, panel efficiency gains mean you need to mount fewer panels.

If 20% more power is produced per panel, you need 5/6 as many panels your project is straight up 16% cheaper.

If anything, your statement about the panels being free and the balance of system and installation cost being expensive make efficiency improvements even more impactful.

The main historical issue with perovskite cells is damage and lifetime, which again directly relate to how many times you need to do these big capital investments. If they last twice as long you need to deal with half as much installation and mounting work over the lifetime.

And labour.. I just got an offer to cover two houses next to each other and the scaffolding is 15k CHF.

This isn't 20% cheaper—it's 20% more efficient. Do they require 20% mounting systems or inverters? If not, you're getting more power without increasing the spending, regardless of which part is the expensive part.

I'd totally believe that more power requires more expensive inverters, but on its own, your comment doesn't make sense.

Mounting system doesn't have to be very expensive. But the mounting labor is. Which is why the mounting hardware does end up costing a lot, because anything to save labor time is usually a good tradeoff.

Inverters are also very cheap nowadays. The most expensive part of a solar installation is labour.

But real estate is not free.

I’d love to swap out my panels for 20% more in the same space if it’s practical.