This useful graph by @hausfath managed to get me angry in early morning. It shows that major models, used by policy makers to assess climate solutions, still wildly overestimate the cost of solar PV systems. The cost they project for *2050* is way higher than the present cost! 
First of all: the IEA WEO line at the bottom is way too high too. As @GJSchaeffer says: “We sell (in EU) at 600 €/kW for anything above 250 kW, and that includes a profit”. Apparently IEA assumed a total cost reduction from 2015 to 2020 of 20 to 25%; a massive underestimation.
The real solar PV capital cost for 2020 is probably around $700/kW, and now I’m being conservative. But the climate solution models have 2020 costs ranging from $1500 (IMAGE) to $2200 (POLES) per kW! That’s extremely ridiculous, trying to avoid stronger language.
I warned for this in 2016, when learned from an insider that such models had present cost of $2,500 per kW, which already was 100% too high at the time. But nothing has changed. Get your act together, climate solution modelers of the world!
Ridiculous errors in present cost of solar PV is of course not a good start to a modeling exercise. But the real damage is of course in the projections. Again, forget about IEA WEO; it has a track record of being wrong about solar PV, and always on the ‘pessimistic’ side.
For *2050*, all climate solution models have solar PV costs *way* above present costs. Trying hard to avoid the use of all-caps here. Think about how stupid and misleading this is.
They range from $950 to $1500 per kW, while just one more halving of the cost would get us to $350!
They range from $950 to $1500 per kW, while just one more halving of the cost would get us to $350!
This is so damaging. These climate solution models are used to inform policy makers of the cost of climate action, but on one of our most important zero-emission energy sources, they get it all wrong, telling the world that it will be expensive. Totally irresponsible.
Of course, it will lead to meaningless future energy mix projections too.
Next time, any modeler working with GCAM, IMAGE, MESSAGE, REMIND, WITCH or POLES wants to present you his/her results on the future energy system, tell him/her to get his/her basic assumptions right first.
Next time, any modeler working with GCAM, IMAGE, MESSAGE, REMIND, WITCH or POLES wants to present you his/her results on the future energy system, tell him/her to get his/her basic assumptions right first.
Next time you read about the cost of the energy transition, and it’s based on such models, take it with a shrug and know and communciate that it will be much cheaper in reality. I doubt that solar PV is the only thing they got wrong.
The data that @Hausfath based his graph on were from a paper by Volker Krey e.a.: Looking under the hood: A comparison of techno-economic assumptions across national and global integrated assessment models. You can find it here: sciencedirect.com/science/articl…. Thanks, Zeke!
Ready to cool down in some healthy February rain now! Blood pressure already dropping ;)
Epilogue: @BloombergNEF's @solar_chase confirms solar PV system costs (EPC) of 600 €/kW for Spain, last year. Thanks, Jenny!
.
.
But: Jenny also states that even lower prices are already possible. A message from @RubenBaetens was now confirmed by a @NavigantEnergy colleague: a total solar PV system cost of €350 per kW was already seen!
That's 380 $/kW in dollars of today, but only 325 $/kW in dollars of 2010, which were used in the graph that this thread started with. I plotted that data point for you.
Now I'm not advocating to enter 325 $(2010)/kW as average cost level for 2020. But anything above 600, that's already 85% above lowest recorded costs, is improbable. And so is any assumption above 300 $(2010)/kW for 2050. In reality, it will be less, I think.
