As we want to get rid of fossil fuels, we will at least have to develop everything now without fossil fuels. And in any case the operational energy. In the case of new buildings this leads to “0-energy buildings”, which generate as much renewable energy as required, locally, and thus take their responsibility in the transition. This mainly concerns solar panels on or at these buildings. Which leaves us with the embodied energy of all materials involved. This is a bit more difficult, but all new construction at least must be CO2 neutral by 2050, or , as in our example, will have compensated the Energy / CO2 of production before 2050, in order to be net CO2 neutral.
Now what does the (simplified) picture look like?
Suppose you build according to the Dutch standard as of 1 January, which is, among others, max. 55 kWh operational building-related energy demand per m2. And suppose such a m2 costs about 5.5 GJ, (1527 kWh) of (fossil) embodied energy to build.
In order to be climate neutral by 2050, first of all the 55 kWh must be generated annually in a renewable manner. But in addition, the embodied energy must also be compensated within those 30 years. That is 1527/30 = approximately 50 kWh per year per m2 extra renewable has to be generated, and only then the house will be a “net” 0-CO2 house from 2050 on.
In other words, 105 kWh / year must be supplied per m2 of floor for the first 30 years (55 + 50). Suppose a Solar panel supplies 157.5 kWh / m2 per year (250 Wp at 1.6 m2), then that would be sufficient to serve 1.5 m2 of floor, or justify its construction within the 2050 target.
However, that’s the optimistic calculation. After all, we also have to include the embodied energy of the production of that PV panel. In a Meta study Khagendra  finds for poly-Si (to about 4000–8000 MJ/m 2 and for mono-Si and 2200–6600 MJ/m 2 [x] . If we use the lowest figure, which is quit optimistic since its without impact from supporting material (BOS), its 611 kWh/m2. That panel will therefore only work for its own compensation during the first 4 years (NL) . The 30 years until 2050 is then reduced to approximately 26 years, and the EE of construction to be compensated becomes not 50 but 59 kWh / year and the required annual production 114, or 1 m2 panel is sufficient to serve for 1.38 m2 of floor.
Again, that is still theoretical. If we calculate a bit more realistically, with a yield of approx. 0.9 kWh / Wp x 157.5 kWh = approx. 140 kWh per m2 panel is produced on average in the Netherlands, and the sum becomes: 140/114 is 1.23 m2 of floor…
Or the other way around, to make a house / building energy-neutral, and to compensate for the embodied energy, in order it to be net climate-neutral by 2050, approximately 0.81 m2 of solar panel is needed per m2 of floor, on or near the building.
Interestingly, it can be deduced from this that with 1 story the roof may be sufficient, but that with more floors, the facade might also be needed. With 2 floors you could maybe manage with a sloping roof with an overhang, which increases the roof surface. This is no longer possible with 3 floors, and the facade will have to be included. That is, if every m2 of the house is heated. If compartmentalization takes place, or only part of the house is heated during the coldest period, then it all works out much easier of course.
Incidentally, it is interesting to note that high-rise buildings can never meet this requirement, unless space is available around the building to install a separate solar panel field. Which rules out high-rise buildings as an argument to achieve higher densities: It no longer applies in a renewable energy world, because you have to take that extra Solar panel space into account in the footprint and density calculation. (It didn’t make sense before as well, by the way, see )
All this is approximately of course, this is a theoretical approach and rough calculation. But it indicates the orders of magnitude we have to deal with.
There are certain bandwidths that apply:The Embodied energy of a m2 of floor can vary from 4-8 GJ, and in special cases even more or less: a clay construction or light straw bale construction can be lower, for example. While if working with aluminum facades, or many installations, you will soon end up much higher.
If we maintain that bandwidth of 4-8, at the practical production value of 140 kWh / m2, then the annual embodied energy to be compensated (in 26 years) varies between 42.5 and 85 kWh, and with that operational energy of 55 kWh that adds up to 97.5 and 140, or 1 m2 of PV, is good for the energy demand anywhere from 1,43 to 1 m2 of floor. Or vice versa, 0.7 to 1 m2 of PV panel is needed per m2 of floor.
Naturally, this variation in embodied energy depends on the materials chosen and, for example, the chosen insulation thickness and other measures to limit the energy demand. A lot has already been invested for 55 kWh per m2 of operational energy demand and the Embodied energy will be on the higher side. When going even lower, to 25 or even 15 kWh / m2, this also has an increasing effect on the embodied energy. For a common construction method and technological energy supply, it can even be 10 GJ / m2.  . One has a positive effect, the other a negative effect: In case 15 kWh / m2 as demand, and 10 GJ or 2770 kWh as embodied energy (107 kWh per year for 26 years), , that equals 122 in total, slightly lower than that 1 m2. But this is guesswork, just to show how it affects each other. Perhaps a 15 kWh per m2 demand can also be realized with a much lower EE. There is a real optimization effort to be made here.
So far we have only included building-related energy for heating and ventilation.
If we also include, for example, 2600 kWh of household energy consumption for a home ( 26 kWh / m2, in the case of a home of 100 m2). Or an additional 18 m2 of Solar panels in total (at 140 kWh / m2).
If we also include an electric car at 10,000 km / year, and suppose consumption at 16 kWh per 100 km, that is 1600 kWh per year. At 140 kWh / m2, that is an additional 11 m2 of solar panels.
In both cases, without solar panel compensation (the embodied energy), or compensation for the embodied energy of the consuming products, such as the electric car.
Obviously, as the national energy supply system will grow to include a larger share of renewable energy, the embodied energy in the production will cause less CO2 emissions. But that is at the expense of an enormous national effort in the production of solar panel fields and wind turbine parks, plus associated infrastructure. which also require (partly fossil) energy. In essence, this impact would have to be (partly) added to the embodied energy of the home, because that infrastructure is (partly) built for that. That is why I prefer to calculate here in energy units, instead of CO2 units. If we were only calculating in CO2, that would only shift our problems towards material impacts: and theoretically imply that we would never have to limit energy use again, as long as it is CO2 free. We just build more and more solar panels and wind turbines.
But in the time required to adjust the national energy supply, which will then consist of more renewable energy, the number of years in which compensation must be made will decrease at the same time: The rough estimates above apply in 2020: in 2021, one year will go off, and over the years the figures become less favorable, in order to be climate neutral before 2050. In other words, more and more “Solar panels” are needed per m2 built floor to achieve the objectives, for buildings that are to be realized later.
There is a rough rule of thumb in all this: approximately 1 m2 of PV is needed per 1 m2 of floor (for building-related energy in new construction, In 2020). Its somewhat exaggerated, but this way we will always on the right side, because in practice it can / will be less, depending on the design, choice of materials, and choice of installation, (think of heat pump). But it will be more , the closer we get to 2050 ….
Disclaimer: Calculations are made on the back of a napkin, and are meant for orientation, to get some feeling about the proportions. All this assuming that solar panels are a good thing, that we have until 2050 to limit CO2, and without looking at the material exhaustion itself. I will come back to these issues.
 Energy payback time (EPBT) and energy return on energy invested (EROI) of solar photovoltaic systems: A systematic review and meta-analysis. Khagendra et all, 2015, Renewable and Sustainable Energy Reviews http://dx.doi.org/10.1016/j.rser.2015.02.057