You are completely ignoring the biggest cost: land. In a dense country like Germany, using millions of existing, "free" rooftops is vastly more efficient and economical than buying and dedicating scarce, expensive land for an industrial-scale farm. Also rooftop solar generates power at the point of consumption. This significantly reduces the need for new, expensive high-voltage transmission lines.
Power consumption in the EU has flatlined and is even falling. New HV transmission lines are not needed - put the solar farms next to some of the shut down nuclear plants (as we now do with big batteries in Australia and decrepit coal plants).
For context, a hectare (10000 m²) directly facing the sun receives 10 megawatts of sunlight at 1kW/m². The city of Kiel is at 54°19' north, so at the equinox a hectare near it receives about cos(54°19') ≈ 58% of that: 5.8 megawatts. If you were to cover the whole field with horizontal solar panels with a mainstream 22% efficiency, the peak electrical power generated that day would be 1.28 megawatts. But you'd need 2.20 peak megawatts of solar modules to do it, because the modules are rated based on directly facing the sun and getting 1kW/m² of sunlight, which they could never do in this situation. According to the prices on https://www.solarserver.de/photovoltaik-preis-pv-modul-preis... those modules are currently 0.100€/Wp, so that's 280 kiloeuros of solar panels. This is about an order of magnitude more expensive than the land under them.
The average produced would only be on the order of 100–200kW, because Germany's capacity factor is for shit. Kiel is in an especially terrible place, with only about 2.7kWh/m² per day on average, according to https://solargis.com/resources/free-maps-and-gis-data?locali....
In real life people tilt the panels in a solar farm toward the sun and leave space between rows of panels so they don't shade each other even in the winter. That is because, even at today's historically record low prices, the panels cost enormously more than the land they're sitting on, so it makes sense to economize a bit on panels even at the cost of needing more land.
This isn't a government failure. The government is correctly pricing the true cost of industrial land use: the loss of farmland and natural space.
Rooftop solar is cost-effective because it cleverly avoids that cost by using land that's already developed. It's not winning because of bureaucracy, but because it's using a limited resource in a smarter way.
That could start happening if Germany's electrical energy consumption were to grow 100× or so, yeah. That might sound like I'm being sarcastic, but no, cheaper forms of energy tend to lead to more energy usage, and solar energy is cheaper, so we might see a lot more energy usage. But we're talking about problems we'll have in 20 years after we've solved global warming.
It happens already. Lease for green land (probably wrong terminus) and arable land went up by more than 30% in ten years [1]. PV is arguably a very small factor of influence yet. But the pressure is real.
That page doesn't say anything about PV. https://www.ise.fraunhofer.de/content/dam/ise/en/documents/p... says that 17 million hectares are used for agriculture in Germany, 48000 hectares are used for golf courses, and (as of 02021) 32000 hectares were used for photovoltaics, which would be 0.2% of the agricultural area. I'm having a hard time finding current figures, but I doubt photovoltaics are more than 2% of the agricultural area yet. I don't think that's why land prices have risen 30%.
Thanks for looking up the numbers, and it wasn’t my intention to exaggerate the influence of PV. I just wanted to point out there is a lot of pressure and it may be a sensible policy to use up rooftops first instead of arable land (even though the influence of PV on land lease rates might be negligible today).
Most people in Germany don't live big and don't have any intentions of doing so. No AC and people have real concerns about the environment and our impact on it.
Obviously it won't happen anywhere from air conditioning. But, for example, atmospheric carbon capture (putting the CO₂ back in the ground) could use a lot of energy.
Most likely, though, it will be things we can't even imagine today.
Imagine that you're in 01980 trying to predict how people will use personal computers in 02000. Would you predict the World-Wide Web, Usenet erotica newsgroups, virtual reality, banner ads, Geocities, MUDs, and spam?
Or, in 01903 trying to predict how people will use flying machines or automobiles in 01923. Would you predict metal airplanes, women getting pilots' licenses, dogfights between forward-firing fighter planes, transatlantic airline flights, strategic bombing from the air, helium airships, and French airmail service to Morocco? Or, would you predict Henry Ford would be making two million Fords a year?
Things change. The last time we had a new cheaper source of energy like this was 250 years ago.
