In the case of Moore's Law, a transistor can be any size. By contrast, solar cells absorbs too little energy if they are too thin.

TL;DR: There are real physical limitations to solar cells.

1. The grams of rare earths required per watt of solar have been dropping steadily. And Solar PV panels with no rare earths at all have been developed, though they are not yet common.

2. "Rare earths" are not so rare. Their prices have tumbled substantially as new mines have opened up and substitutes have been found.

Yes, but mining them is an environmental disaster.

Yeah, but the super-efficient solar cells are not the cheap ones that everybody is putting on their roof. They're the ones going on the international space station.

>Once a particular solar technology reaches grid parity

It already did in about half of the continental US, about two years ago.

I mean that seriously. Right now, pretty much all that matters is cost per watt. For normal terrestrial uses, a 20% efficient cell loses out to a 5% efficient cell that's 10x cheaper. (Numbers made up on the fly with no particular basis.) We're not currently hurting for space for these things, and sunlight is free.

But solar thermal has the advantage of integrated storage. When solar overall is at some reasonably high penetration (say, 10-15%) storage will become important. At that point, it depends on the cost curve of storage technologies.

If batteries or other energy storage technologies have gotten cheap enough, PV + storage will be cheaper than solar thermal. If batteries haven't dropped in price enough, solar thermal may see a resurgence.

My bet is on PV + storage, though there is no certainty here.

Of course, thermal generators required direct incidence (don't work well on diffuse light), but the long and short of it is that with thermal we (may) need about half the total area we would need for PV. (On the other hand, PV can practically uses space - rooftops - which is inaccessible for thermal.)

Nice blog, by the way :-)

I know they're different technologies, but I wonder on the cross-over effects of the two. If people see a solar thermal plant going up near their town, will it encourage or discourage them from installing PVCs? Are nation states that encourage / subsidise large-scale solar thermal deployments more likely to encourage end-user scale (PVC-based) systems?

Entirely agree managing storage is increasingly the pain point - in both scales of systems (thermal just rides out cloudy / dark periods more smoothly, I believe).

Is it still true that the most $-effective large-scale power-storage mechanism is to pump water up a hill?

I don't think this will make much difference - individual PV installation is an investment decision (just like other big home improvements) based on cost of energy and subsidies. It makes some difference to how "green" people feel, but the cost matters more. It may have an effect in planning battles. Here in the UK (esp Scotland) it's all about the wind farms. European solar thermal is currently struggling against low natural gas prices: http://www.estif.org/statistics/st_markets_in_europe_2014/

Pumped storage requires naturally occuring suitable locations (not too many available) and incurs environmental damage of damming a river.

(Calling it "PVC" is really confusing me with the flexible polymer)

As to influence - I think HN types are more informed than the GP, so in our bubble it may be hard to speculate accurately on any such influence. By which I mean, if you're not informed about the technology or the trends, then any intrusion of these technologies into your life will be more significant. Perhaps I am wrong.

For Scotland I would think we're some way off a sufficiently efficient cell, given hours and angles of incidence of sunlight.

Disclaimer - I'm in Australia, about 35 degrees south, and to be candid don't really understand how even places like Germany have so much love for PVCs ;)

How do you think the fact that natural gas is guaranteed to decrease in availability / increase in cost (over the long term) compared to the guarantee solar will decrease in cost, will influence users? And how do you think the idea of users being able, or being primed, to think of themselves as producers as well as consumers will influence their decisions on power?

I don't think individuals are very good at adapting to the long term market. If I know gas will be more expensive in 5 years, I can't do very much with that today. There's also the old "the market can stay irrational longer than you stay solvent": if I invest too early I can just as easily lose money.

In the UK, I think the general public are ahead of the media and politicians on the issue. There's a noisy anti-environmentalist and anti-windfarm contingent, but most people recognise the tradeoffs. I don't think people want to be their own producer en masse, any more than grow their own food or be their own bank, but if they can make a capital investment that saves money and produces income then you're speaking the right language to the middle class.

People might start to get vocal when gas prices go up in the UK; complaints about fuel poverty and demands for more fracking. It'll require a long period of gas being more expensive than electricity for fuel and heating for people to switch their homes over.

How serious are grid stability issues presented by solar PV (without storage) at high penetration rates (say > ~ 20%)? Important? Not? Thanks.

They are not trivial. Germany claims one trillion EUR direct costs on network upgrades to cope with renewable energy. They do not take into account energy storage costs. Even now (at least in my country) 75% of energy costs are distribution costs.

Energy storage is indeed going to be vital once solar exceeds 10-20% of capacity. Energy storage, fortunately, is also dropping fast in cost:http://rameznaam.com/2015/04/14/energy-storage-about-to-get-...

I have future posts coming with closer looks at the future of storage prices, wind prices, and possibly one on nuclear.

http://www.world-nuclear-news.org/NP-Trillion-Euro_cost_of_G...

I do agree that Moore's Law generally holds true. However a better model is an approximation of multiple S-curves, where each S curve represents the beginning, growth, and maturation of each step in technology.

The smearing of S curves and integration of new lines of tech allow for what appears an exponential growth.

Also, I'd like your opinion on http://www.bioinspired.net/ -- Commercially available memristors, albeit high priced.

Its likely it'll never be cheaper to build traditional sealed glass panels cheaper than house windows, although per sq meter it might get cheaper than televisions.

Some of the more optimistic claims outside the article seem to have very unusual assumptions about the cost of glass.

So since the cost drivers changing - why does it make sense to model future price on old cost drivers ?

I initially set out to model soft costs on their own, separate from module costs. However, that effort revealed other factors. Solar capacity factor is rising and operational costs (which are not included in soft costs are dropping. There are more variables at play. For some of those variables, there isn't robust data available over lon times.

All of those variables ultimately feed into electricity prices, in the end, which is why I chose that modeling approach.

Even if these two use the same basic technology, I think they are essentially two completely separate "markets."

The inverters used in the residential market versus the solar plant market are usually quiet different and, IMHO, form two distinct markets. Reductions in the inverter cost for solar plant deployments would result in a reduction in the home market, but the relationship is not as direct as the relationship between the panels since you would rarely be able to use the same inverter in both markets.

I.e. I'm trying to figure out if going solar now is better (in California) or just waiting a few more years. I'm in no particular hurry since my electric bill vs. sq/ft is already pretty good.

Rooftop solar will probably also take 2-3 years to make up for the cost lost to the subsidy reduction. It may even be 4. So after 2016, I'd expect rooftop prices to rise before slowly dropping, and getting back to 2016 levels around 2019 or 2020.