April 29, 2009
Solar Photovoltaics Prices Coming Down

The Q1 2009 conference call between photovoltaics maker SunPower's top executives and financial analysts covered the question of how fast solar panel prices are falling. Answer: pretty fast. While SunPower claims to have cut average selling prices by less than 10% so far this year the analysts point to competitors who've made bigger cuts and the CEO of SunPower admits they see another 20% of price cuts coming in the rest of 2009.

Nicholas Allen - Morgan Stanley

So you expect another 20% decline in pricing over the course of the year?

Thomas Werner

Yes. The way you should think of it from our perspective is that we are prepared for... or we've tested into our models up to 20% more price decline this year.

SunPower is probably the second strong player in the PV space after First Solar. This statement by its CEO is therefore very telling.

If you are thinking about putting solar panels in your house then wait. You can't possibly make or save enough money off those panels in the next 8 months to return you 20% of your costs.

The reference here to "poly pricing" is for polysilicon crystal which is a production input to silicon PV production (though not for thin film PV as made by First Solar). I've read other sources claiming that prices for poly have declined from a peak of $450 to $100 per kilogram.

Pavel Molchanov - Raymond James

Target for a 50% all-in cost reduction by 2012. Given that poly pricing had collapsed much faster than pretty much anyone would have expected, do you see any upside to that target?

Howard Wenger

Thanks Pavel for the question. Tom, I'll take this to keep us moving quickly. We also guided that two-thirds of that cost reduction which happened by 2010 or sooner. And yes, we are seeing some opportunity for that to accelerate.

So SunPower expects to cut their own production costs by a third by 2010 and more by 2012. This tells you where solar PV costs are going and how fast they are going there.

For the several years I've been writing posts about the prospects for solar panels a core of skeptical commenters have consistently reacted by claiming that PV costs have declined so little for so long that PV is never going to become viable. But look at the numbers above. The era of relative price stability in PV prices has ended. The era of rapid price declines is upon us. It is about time.

Share |      Randall Parker, 2009 April 29 10:28 PM  Energy Solar


Comments
K said at April 30, 2009 1:57 PM:

Randall: Prices move down for reasons besides lower production costs. But the driver this time may be lower costs. I'll leave that to those who want to figure it out.

Prices move down on falling demand. Probably not the case here. But they also move down when supplies temporarily exceed demands. And that may be the case; companies all over the world have been rushing into polysilicon production for about three years. And into PV factories too. Supply may have grown faster than demand.

Demand, and hence prices, can also move on anticipation. Your suggestion, postpone buying now because prices will fall, is based on anticipation. Subsidies can play a large part in anticipation. What will the government add to or phase out in rebates? And in tax credits?

Large LCD television prices have fallen a lot. It that from oversupply or from recession? It certainly isn't because people don't like them.

And the market share of hybrid cars hasn't been sparkling either. From anticipation, fuel costs, or recession? I suspect postponement. Many new models are due within a year but few are in showrooms now.

John Mooree said at April 30, 2009 7:42 PM:

Installed system cost is a more interesting metric.

Randall Parker said at April 30, 2009 8:28 PM:

K, Prices are dropping due to both supply and demand reasons. Lots of PV companies started up since 2000. They've been ramping their production capacity. At least some of them have found ways to cut production costs significantly. But demand rose so fast until the second half of 2008 that prices stayed firm and that encouraged even more building of production capacity.

Finally demand plummeted for a few reasons. The financial crisis cut financing. Demand for electricity dropped. Cost of fossil fuels for generating electricity dropped. Suddenly the PV industry had a huge excess of production capacity and they had to cut prices.

I do not expect prices to recover. This decline is one way. By some time in 2010 you'll be able to get PV modules for a quarter off current prices.

John,

Yes, installation costs are substantial. But the installers have got to be under pricing pressures too.

Paul F. Dietz said at April 30, 2009 8:54 PM:

First Solar says they have manufacturing costs for CdTe modules (this includes everything, including raw materials) under $1/W, and say it will be $.63/W by 2012.

The limit on how many CdTe modules can be produced is ultimately set by the availability of tellurium, but the price of Te could rise a lot before it becomes a significant part of the cost of a CdTe module. Huge amounts of Te are in tellurium-rich ferromanganese crusts in the deep ocean.

Wolf-Dog said at April 30, 2009 11:42 PM:

I am worried about the Cadmium that First Solar is using. If the world starts using these Cadmium based cells everywhere, some of it will leak and pollute the environment.

Paul F. Dietz said at May 1, 2009 8:57 AM:

I am worried about the Cadmium that First Solar is using. If the world starts using these Cadmium based cells everywhere, some of it will leak and pollute the environment.

It's a very small amount of cadmium, since the semiconductor layer in CdTe cells is maybe 3 microns thick. Moreover, the world has a surplus of cadmium production (from zinc/copper/etc. refining), so the element is being liberated from ores anyway.

Nick G said at May 1, 2009 2:08 PM:

Paul, could you elaborate on tellurium availability?

How much Te is in a sq meter of CdTe cell? What's current and historical pricing? How well demonstrated is the availability and feasibility of deep ocean tellurium mining?

D6man said at May 1, 2009 2:37 PM:

SolarPlaza did a thorough 3 year analysis of supply and demand of PV in July 2008. Supply was built up by accumulating expansion actions by module makers and silicon refiners. Demand was built up by adding country by country demand, taking into account ranges of possible government incentive programs. Bottom line is they forecast global supply rising 70% per year and global demand rising 30% per year through 2011. I should have sold my solar stocks then. :(
Obviously this won't roll out completely according to plan. The marginal, smaller, and higher cost suppliers won't make it. But the lower cost or better positioned like First Solar and SunPower will. Prices are being led down by the lower quality, mostly chinese suppliers. But most of these don't make a product you'd really trust on your roof for the 25 year warranty period.
Up until late 2008, solar module costs were about half of total install costs. As module prices fall, state incentives are already falling too, so I'm not sure waiting will necessarily result in a lower cost to the homeowner. Also, there is the dynamic of state funding programs running out of money, which also argues for not waiting.

Eric said at May 1, 2009 4:34 PM:

For the several years I've been writing posts about the prospects for solar panels a core of skeptical commenters have consistently reacted by claiming that PV costs have declined so little for so long that PV is never going to become viable. But look at the numbers above. The era of relative price stability in PV prices has ended. The era of rapid price declines is upon us. It is about time.

The companies making silicon crystal have been enjoying enormous margins over the last few years, and that was destined to end either as the demand for chips fell due to recession or the entrance of new producers into the market. We'll probably have a glut for a few years until the demand for integrated circuits recovers. I agree - solar cell prices are in for a precipitous decline.

Randall Parker said at May 1, 2009 6:52 PM:

D6man, Yes I've also read that the split between module costs and other costs was about 50:50 for houses. But what about for large desert PV farms of the sort that electric utilities are building? I'd expect a much larger fraction of the total cost there would be the modules.

Also, if the PV is getting put on a house during original house construction I'd expect a lower cost since the construction workers don't have to tear into the house to install the PV equipment. They can design and build it in. So I would how much one saves on installation cost if PV is part of the original design.

I'd like to know how much of the non-module cost is due to labor costs, electrical conversion equipment, or other costs. What are the prospects for lowering each type of cost?

Also, when PV has higher conversion efficiency does that lower installation costs?

Chris said at May 2, 2009 4:33 AM:

Randall, the PV panel costs are dropping - which is great - but not the grid-tie inverter costs. If you want set up a PV GTI system with reasonably serious capacity it costs an unusually large amount of cash for the inverter given that it is really just a box full of electronics.

Randall Parker said at May 2, 2009 8:26 AM:

Chris, Any idea what's in a grid-tie inverter? Are the parts already produced in high volume for other uses? What's the mark-up on a GTI box as compared to its parts cost?

I'd expect in a recession there'd be pricing pressures on GTI boxes.

I'm looking at grid-tie inverter costs on Google Shopping and see lots of prices in the hundreds and thousands of dollars. Any idea what capacity grid-tie inverter one needs for a typical house? I'm trying to get a sense of what percentage of a solar install's cost comes from the GTI.

Paul F. Dietz said at May 3, 2009 6:21 AM:

Paul, could you elaborate on tellurium availability?

How much Te is in a sq meter of CdTe cell? What's current and historical pricing? How well demonstrated is the availability and feasibility of deep ocean tellurium mining?

CdTe has a density of 5.86 g/cm^3. It contains 3.12 g/cm^3 of tellurium. If the CdTe layer is 3 microns thick, a 1 m^2 solar panel will have about 10 grams of Te. If the panel produces 100 W peak (around 10% efficiency), this is about 100 mg/peak watt. The cells can likely be made thinner if Te becomes more expensive, since the absorption length of the relevant wavelengths of light in CdTe is much less than 3 microns.

The price of Te in 2008 was $215/kg (or about $.02/peak watt of PV capacity). This was higher than previously, in part due to increased demand from CdTe PV manufacture.

World annual production of Te is not currently available, but doesn't appear to be more than several hundred tons. Reserves are 22,000 metric tons, with a reserve base of 48,000 metric tons. The figures there are assuming only half the Te in copper anodes slimes is recovered. This does not include seafloor deposits, for which the technology for extracting is not yet proven. A metric ton would yield about 10 MW of peak PV capacity.

http://minerals.usgs.gov/minerals/pubs/commodity/selenium/mcs-2009-tellu.pdf

th said at May 3, 2009 6:32 AM:

If low demand and huge subsidies keep up, you might be able to get your 4kw system for free, thats obamanomics.

Randall Parker said at May 3, 2009 9:32 AM:

Paul, At 100 mg/peak watt (100 grams/peak kw) if the 22,000 metric tons (22,000,000 kg) were all converted into PV then the result would be 220,000,000 peak kw or 220 peak gigawatts. Is that correct? That falls way short of what one would hope to achieve with a heavily PV world.

What's the effective average percentage of the peak output? It obviously depends on where the PV is located. Suppose it is all in Arizona. Do we get a third on average over the year?

Paul F. Dietz said at May 3, 2009 10:09 AM:

In desert areas I believe the average output is about 25% of peak, for fixed flat plate collectors without mirrors. You could boost that a bit with fixed mirrors or other non-imaging optical systems; I think CdTe tolerates higher temperatures than x-Si before efficiency declines.

I've read that Te is actually, in the universe, the most abundant element with atomic number >= 40. Its rarity here at the Earth's surface must be due to some chemical fractionation; I think most of it may be down in the core. This suggest it might be mined in space. There had also been a notion that the radar-reflective highland areas on the surface of Venus were due to tellurium vapor-deposited there (since they are somewhat cooler than the lower areas), but that's now thought to be due to vapor deposition of lead (and perhaps bismuth) sulfide.

Randall Parker said at May 3, 2009 10:46 AM:

Paul, I do not foresee much use of solar concentrators to focus light on low efficiency thin film PV. If one is going to go to the expense of focusing light from a larger area onto a smaller area the cost of the PV becomes such a small portion of total cost that it makes sense to use the highest efficiency silicon. So buy from SunPower, not First Solar.

So far concentrating solar is mostly used for steam. Will that picture change? Will focusing ever make economic sense in a PV-based method of moving the electrons?

Paul F. Dietz said at May 3, 2009 10:57 AM:

Randall: low concentration systems need not be expensive. They do not require tracking, for example.

Systems making steam are high concentration systems, which are another kettle of fish entirely.

D6man said at May 4, 2009 6:51 AM:

Randall, to answer your question on residential inverter size, a 4 kw (AC) inverter coupled with 4 kw (DC) of PV would generate about 4.5 to 5.5 Mwh per year, depending on where you live and how well the array is sited. That's about the average residential annual load. Residential inverters cost about 60 cents per watt of AC rated power. Competition in inverters has heated up in recent years but prices haven't moved much yet. Balance of system material can be another 60 cents per watt (wire, conduit, PV module mounting hardware, circuit disconnect and protection equipment, etc.) Install labor and benefits (fully loaded) is another 60 cents per watt. Then add in permitting, inspection, engineering, incentive and utility paperwork costs.
Larger commercial and utility sized systems have also been about 50:50 modules to all other costs. But in 2008 when residential systems averaged around $8.00 per DC watt, the big ones were around $6.50.

Post a comment
Comments:
Name (not anon or anonymous):
Email Address:
URL:
Remember info?

                       
Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright