There is an interesting article in the current issue of Discover magazine for a device that uses a bunch of mirrors (96 partial parabolic mirrors) to focus sunlight on the base of a Stirling engine. Light sensing photocells are hooked to thermocouples and a computer to keep the mirrors focusing the sunlight on the engine, which turns a generator which produces the electricity. The heat focus on the Stirling is about 1,200 degrees Fahrenheit. The prototypes have been built and work well. It looks very interesting. In mass production they think they can sell a unit for about $250 bucks, and it produces 250 peak watts of power. They get that $250 per unit figure by assuming a production run of 100,000 units. Target customers will probably be commercial customers with flat roofs (think light industrial or office parks) in the deep South. Of course it won't work as well on cloudy days etc. but there is nothing fancy here, no special technology. It looks like they can produce electricity at $1 per watt or a little more (assuming they can mass produce the units) which is reasonably competitive.
My guess is that this venture will be successful, but I wonder about maintence costs as well since these things do have moving parts. The design looks good (and you should be able to read it online in a couple weeks, Discover doesn't post the online issue until after the print issue has been out a while and this is the August 2003 issue). I bet that they will need maintenance every year or 6 months, at a miniumum to wipe of the glass enclosure that protects the unit from the weather but probably also to check the moving parts.
dark_lord says: "...
you should be able to read it online in a couple weeks..."
Sounds interesting, but if the article is speaking of the Kamen Stirling engine, it is only 20% efficient at converting solar energy to electricity and at the expense of having a large mirror array that must be constantly moved about to focus the light on the engine. Given the area needed for this generator, I'm not sure it would represent an improvement over current PV cell technology and almost assuredly the system cost will not be anywhere close to $1/W. But nevertheless, please post the article when it comes online. It should be an interesting read.