Monday, November 30, 2009

Supply (we got too much) and Demand (ain't got enough)

There may be nothing new under the sun, but the cost of solar electric is falling rapidly due to an oversupply of polysilicon. With new foundries coming on line in China this is something that will not be temporary. The estimates above are from November 2008, but the fall has been even steeper, having reached $60/kg today. At about $40/kg solar becomes competitive with fossil fuels on costs.

On the other hand, a low price for polysilicon is not everything. With the economic slowdown, some of the older polysilicon producers are getting priced out of the market, and solar modules are also moving slowly. Production is leaving the US and moving to China. GE and BP are closing plants in the US.

Comments?





Here is the beginning of my post. And here is the rest of it.

9 comments:

John said...

Eli, can you provide some mor information about what it takes for solar to be competitive with fossil fuels? I had the impression that the "balance of system" costs (i.e., everything besides the PV module) was also a significant cost problem.

-John

Anonymous said...

Actually, no comment, but the atmosphere in the US clearly isn't pro-renewables, since they're pricing themselves off from the fastest growing market. Meanwhile, some Tienshan village maybe getting it's first electricity from a clean source, since some distant relative of a local tycoon has moved in there to get away from the pollution in nearby city.

John Mashey said...

1) This is just another example of typical boom and bust cycles in commodity semiconductors. Back when I was a microprocessor guy, we used to see this on the DRAM business all the time. DRAM prices up, bunch of people all decide independently to build new fabs. Then, oversupply.

2) This one has additional twists. For a long time, solar cells were in effect made with "extra" silicon from semiconductor production, which has much more stringent specs than solar grade.

Then, the big rampup of solar demand exceeded the supply, at which point a bunch of people decided to build more efficient production facilities geared for solar grade.
Hence, transient oversupply.

3) Balance of System costs always matter, but people working very hard on those as well. BoS costs are especially lowered for be contruction, for integrated PV tiles, and with some better inverters.

4) Silicon costs still matter. As it happens, I sometimes review solar startups for VC friends. Without giving anything away, there are a bunch of techiques coming along that should provide godd cost reductions, in some cases greatly lessening the use of polysilicon.
I don't know which ones will actually work, but for the first time I can recall, Silicon Valley is seriously spun-up on thus turf, and some of these techniques will work.

Anonymous said...

Polysilicon is one of the things China has recently been accused of “dumping” on the world market.

“The State Council, China’s cabinet, recently singled out the iron and steel, cement, electrolytic aluminum, glass, coal, chemical, polysilicon and wind power equipment sectors as the worst offenders when it came to overcapacity and announced steps to rein in their expansion…

…The State Council, China’s cabinet, recently singled out the iron and steel, cement, electrolytic aluminum, glass, coal, chemical, polysilicon and wind power equipment sectors as the worst offenders when it came to overcapacity and announced steps to rein in their expansion…”

http://www.nytimes.com/2009/11/27/business/global/27trade.html

arch stanton

Anonymous said...

Here's a report that suggests that "green" costs jobs. The implication I suppose is that greenies are out to get Mr Jo Main Street. I suspect we'll be hearing a lot more on this one.

http://www.juandemariana.org/pdf/090327-employment-public-aid-renewable.pdf


Mike Donald

Anonymous said...

Crine out loud Mike keep up with the conversation. Georgie Will span this juan months ago.

http://www.washingtonpost.com/wp-dyn/content/article/2009/06/24/AR2009062403012.html

Would have been wittier typing Mr. Juan Main Street. But from what I heard on the news solar panels will be grid comparable in say 10 years. Jobs a goodun.

Mike Donald

EliRabett said...

John, Eli is very amateur at this stuff. The links may provide you a better answer, which is one of the advantages of the Internet.

Frankly, this looked interesting as a big downward step in the cost of solar. Yes, there are other issues, and yes there are other flavors of solar, but the point is that the technology is not static.

John Mashey said...

For sure, there are many reasons to predict continued downward pressure on silicon prices, long-term.

let me try again:
a) Some of this is the same kind of trend of technology volume-cost scaling people are used to in the semiconductor business. That's a continuing process.

b) Some of it is the shift from semiconductor-grade to solar-grade. Some of that is one-time (say like going from discrete transistors to integrated circuits), and some is more of a).

c) And some is just long-familiar noise of overbuild/underbuild boom-and-bust, overlaid on the trends above.

That is, as usual, one must understand the "physics" behind the curves. :-)

BUT, there is no doubt that solar PV is going to keep getting cheaper for a good while.

Marion Delgado said...

Just one. We have a solar plant in Eugene, OR. But it's owned by a Chinese firm (mainland) and most of the money goes back there. Still, some of the jobs are local, and the subsidies from the city are less than the Korean microchip plant that shut down was getting.