Yes. Not indefinitely of course but we've seen nothing yet and there's enough in the research pipeline to suggest clear trends over the next two decades in terms of cost improvements, efficiency improvements, cheaper materials, etc. That 4% you mention was less than 1% not so long ago and will be closer to 40% not too long from now. Such is the nature of exponentials.

IMHO 5X cost reduction is basically a done deal even just applying simple economies of scale. You can argue whether this happens in five or ten years maybe but arguing it isn't happening seems futile. Beyond that, 10x is extremely likely to happen as well. There's a lot of research happening that would need to be a complete failure/waste of time for that to not happen. At worst it will happen slower that I might like. Maybe this happens over the next decade. Or two. Or three even. From there, 20x might be quite possible as well.

Beyond that we're indeed talking rapidly diminishing returns. 10X would mean 0.2 cent/kwh. 20x would cut that to about 0.1 cent. This indeed gets you in to territory where metering it becomes more expensive than is worth the trouble. The average household uses about 15000 kwh per year, or about 150$ worth of energy at these prices; much less outside the US. Much of that is going to be produced on people's roofs or in their back yards. At that point it turns into a fixed cost.

Historically the same sources producing the studies you are referencing have been off by magnitudes predicting current adoption and prices for both solar and wind. So, I'd consider that the pessimist glass half empty point of view. Not to dismiss it entirely but arguing cost increases seems a bit far fetched in light of current trends in the market.

Aren't roof solar panels very inefficient? They can't track the sun and you can't have good density from them either. Furthermore, they don't work for office buildings and apartment buildings.

>Not to dismiss it entirely but arguing cost increases seems a bit far fetched in light of current trends in the market.

I think what the parent was saying is that because solar and wind don't reliably produce energy, we will need redundancy. As the market share of solar and wind grows we will need to increase the amount of storage and power generation far beyond what we normally need to operate.

Let's say you have a city that uses 1 GWh of electricity per day. If you were to generate all the electricity with natural gas then you need to have 1 GWh plus some extra for redundancy to fulfill the needs of the city. If you generate 5% with solar and the rest with natural gas then nothing really changes. If the weather is bad then the natural gas plant will step in.

As the share of your solar power generation grows toil need more and more redundant storage and solar power generation. If you rely 100% on solar power generation and the weather can be bad for a week in a row over the city, then you'll need storage that will last for more than that entire week. Not only that, you'll also need extra power generation capability to be able to fill and maintain that storage.

Just to put this into perspective: the US uses about 10 TWh of electricity per day.

> Aren't roof solar panels very inefficient?

They are good enough. Low efficiency just means we have to buy more of them. Obviously we'd need solar plants (and maybe wind parks and other clean sources) to also power those with limited access to roof surface.

Of course, bad weather doesn't mean solar stops working; it just reduces the output. The effect is typically very local as well. So, all that means is that you need a bit extra capacity to cover for that or import energy from somewhere else where the weather isn't miserable.

When (not if) the price for solar and wind drops 10x, you'll be able to buy 10x more than what you need and still beat coal/gas/nuclear on price. 10x is an insane safety margin that would mean you produce more than you need even on the most cold, dark, and miserable day imaginable. And of course solar is not the only source of clean cheap energy. Wind would be another popular option.

Bad weather in northern Europe means that everything is covered in snow for a long time. I think that's slightly more than just reduced output. The snow also tends to cover am area that's hundreds of kilometers in every direction too.

I've lived in Sweden and Finland. It's actually the lack of sunlight that is causing issues with solar. The snow is much less of a concern. E.g. Helsinki only gets a few hours of daylight and the sun does not get much above the horizon in the winter months. The reverse is true as well; you get insane amounts of sunlight in the summer months. Luckily there's wind, hydro, wood pellets, and a few other alternatives.

Further south in Germany, Netherlands, etc. Solar is usable throughout the year and quite common. Obviously output in the summer is going to be much better than during the winter.

> Aren't roof solar panels very inefficient? They can't track the sun

We are well past the point where tracking the sun was desirable. The apparatus that allows solar panels to track the sun is now much more expensive, especially when you consider the maintenance costs, than just using more static panels.

Actually, most new utility scale PV in the US southwest is now using single axis tracking systems. The advantage here is that the power curve is flattened over the day, which both makes better use of the inverters and pushes more production into higher value time periods.

With dirt-cheap renewables it will be more efficient to generate methanol from solar or wind electricity and to use it for long-term storage and redundancy than to burn natural gas or coal. And compared with coal or natural gas with methanol one can use fuel cells to convert it to electricity so one can have relatively small plants for backups in case of really long strips of bad weather.

This is very very optimistic. While the prices of panels are going down the rest of the cost isn't (land, permitting, installation labour, etc). And panels are only a small proportion of the cost now.

There is also only so much more efficient the panels can get per sq m. Right now we are approaching 20% efficiency, so the idea that they are going to get 10* cheaper with complete ease seems to be pushing it.

Regardless solar and wind have a lot of externalities that aren't really priced into these 2c/kWh PPA prices. If you need to build a fleet of natural gas plants that are only operational 10% of the time and therefore cost 50c/kWh to take up the slack from low solar days it needs to be priced in somehow.

All the costs have been falling, at various rates.

It's amusing to see "permitting costs" being presented as a barrier, when at the same time "get rid of regulations" is spouted as the panacea for reducing nuclear's costs.

> That 4% you mention was less than 1% not so long ago and will be closer to 40% not too long from now. Such is the nature of exponentials.

But such is not the nature of the logistic function, which is usually a more accurate model of how technologies penetrate. (yes, in the beginning it looks like an exponential)

There is a final cost decline "baked in" to the curve, though. That's because the lifespan of solar now for calculations is maybe 20 years. Anything more would not make sense as costs are declining so fast. When the tech matures, the final part is extending the lifespan to 40-60 years.