What would it take for solar power to move out of the niches to which it's been confined and start to compete directly as a primary source of energy? Cost is one of the biggest factors, and several new technologies for solar collectors offer the prospect of significant reductions, as described in this article from the San Francisco Chronicle. But the cost of collectors is not the only obstacle solar must overcome. Like wind, solar power is an intermittent source, and this must be factored into how and where it can be used. Even with the potential for lower unit costs described in the article, solar has a ways to go to compete on a level playing field with electricity derived from fossil fuels.
It's also not clear which is the most relevant cost on which to focus. The above article compares only capacity costs (though this is never clearly stated) that relate to the expense of building and installing a power plant or solar array. These costs are measured in dollars per kilowatt (kW) of capacity. This is a poor basis for an apples-to-apples comparison, however. The market tends to look at the price of flowing electricity at various points in the distribution system, measured in cents per kilowatt-hour. Consumers ultimately pay the retail price, while generators and traders deal with various levels of wholesale pricing. Solar is hard to compare on a flow basis, since its costs are essentially all capacity-related, with a negligible ongoing cost. Its pricing on a delivered unit of electricity basis depends on many assumptions, particularly with regard to financing.
For example, a 2 kilowatt solar rooftop array for consumers would cost $2,000, based on the most optimistic figures ($1,000/kW) from the article and ignoring costs of installation, DC/AC conversion, and other factors that would increase the real-world installed cost significantly. This also ignores any tax credits that might be available. In a favorable location such as Southern California, such a system would collect on average 5.5 peak sun hours per day, generating about 3200 kW-hours over the course of the year, after wiring and inverter losses and sun-angle factors. If the cost of the solar panels is amortized over ten years at 6%, this translates to an effective cost of electricity of 8.5 cents per kW-hr.
This figure would be competitive with retail power costs in most of the country, but not with wholesale costs or the prices many large businesses pay. In other words, while a roughly five-fold cost reduction versus the current technology would position solar power nicely in the market for home power--a sizeable market to be sure--it still misses most of the business/industrial market, even ignoring the substantial added costs of turning its intermittent output into a continuous, reliable power source by adding storage (batteries or ultracapacitors) or generating hydrogen for use in fuel cells.
When fully-deployed, mass-market rooftop solar power would have a major impact on utility planning for peak generating capacity needs, without affecting baseload power demand much. Paradoxically, then, this puts rooftop solar in competition, not with coal-fired or nuclear power plants, but with gas-turbine plants that are already among the most efficient and environmentally-benign energy assets out there. In effect, cheap solar panels are really a way to reduce natural gas consumption in the electricity sector, freeing it up for other uses (or reducing future import requirements.) I wonder if this is what most solar advocates have in mind?