Saturday, February 16, 2008
Will the much-anticipated "green power" revolution come from massive government programs to develop new technology top-down, or from the venture-backed companies built on the foundation of America's semiconductor industry? Here is one answer.
I'm with T.J. on this, as I almost always am.
If it is true, in 8 years the cost of photovoltaic power wil be about16x less than it is today and 20 years from now it will be 1024 times less (finally, electricity too cheap to meter!).
Please explain how that will occur.
Moore's Law concerns the increase in the number of transitors that can be fabricated per unit area. How does that relate to the cost of photovoltaics?
To expand on the comment of Anonymous above, the limiting factor of collecting solar energy is the amount of solar radiation per square meter available on the surface of the earth where the solar panel is installed. It is night half the hours in the year everywhere and, except for desert areas or extremely high mountains, often cloudy. More transistors per square mm might raise the collection efficiency, but won't bring more sunlight to the surface or drive away the clouds.
As to the free market hokum, Germany is currently installing lots of solar panels (hence the location of the factory). What free market fans ought to know is that Germany mandates that power from solar generators be purchased by grid operators at an astoundingly high, government-fixed rate per kwh. The cost of this generation is, of course, passed by law directly to the consumer by the German Big 4 grid operators. There would be few solar panels installed here without government distorting the "free market" for which T.J. professes such a fondness.
Traditional architecture solar panels are a technical dead end and the cost won't drop much in the future because the cost of a fab has to be amortized.
The next solar breakthrough will be something completely different with production costs that are dramatically lower.
Moore's Law is driven by market forces and reliable expectations for continued technology improvement. Solar cells have a theoretical upper limit which is approached as an asymptote. Maybe the distance from the asymptote can be halved every 5 years or something like that, but that improvement will quickly reach a point of diminishing financial returns and people will lose interest.
If you look at alternative energy as a whole, then maybe Moore's Law applies. There are so many ideas on how to tap a variety of energy sources, and so many ideas about innovative ways to use it efficiently, that it could happen. But once again, you need the market incentives. Nothing is really going to happen unless we find artificial ways to keep fossil fuels expensive, or at least more expensive than some of the potential alternatives. I'm predicting that the Singularity will not arrive until after Peak Oil, basically because we won't care enough to make the necessary investment.
Once we put collectors in space, then solar energy could begin its true exponential phase. Right now, it's just a logistic curve that will top out at some point. Nuclear has a much more promising future as of today, basically because we can make it cost competitive even without externality charges.
Would it not be wonderful if we had started building several new reactors twenty years ago?
Not only would we have them operating today, but the cost would have been a fraction of what it would cost now.
Oh...there is an idea, if we start now, the cost will be lower than twenty years in the future.
When we will really have to have them.
Maybe..if we are around to operate them.