Like most people in the solar industry, the first question that popped into my mind was: is $1/W realistic? I will try to answer that question for residential solar. Since the DoE's SunShot initiative is focusing on (centralized) utility-scale solar, we'll need to convert the goal of cost competitiveness into fully installed cost for (distributed) residential roof-top systems. We know that residential electricity costs on average $0.12/kWh in the US today (see earlier post The Cost of Solar Electricity). Assuming an inflation rate of 2%, that will increase to $0.14/kWh by 2020. Taking a financing rate of 5%, installed costs for a residential roof-top system would have to decrease to $2.50/W by the end of the decade for solar to compete with distributed electricity. Given that currently the cost to install an average sized residential system costs about $6/W, this is an equally ambitious goal as getting utility-scale solar to $1/W. So can it be done?
Let's first break down the current cost structure for a standard 5kW residential system:
$/W Total $ Possible Cost Structure
Modules: $2.00/W $10,000 $0.50/W (1)
Balance of System: $1.00/W $5,000 $0.50/W (2)
Project Management: $1.00/W $5,000 $0.50/W (3)
Permitting Costs: $0.50/W $2,500 $0.25/W (4)
SG&A $1.50/W $5,000 $0.75/W (5)
$6.00/W $30,000 $2.50/W
Thoughts on the possible cost structure:
(1) The DoE hopes for module prices to come down to $0.50/W by 2020. We have seen prices decrease significantly over the last years and further reductions are expected especially as manufacturers build out capacity and optimize manufacturing processes. First Solar is already at $0.75/W manufacturing cost. Conclusion: $0.50/W is possible.
(2) Similarly to module pricing, there is significant potential for balance of system costs to decrease. Especially inverters should come down in costs as competitors enter the market and the large players scale up production. Conclusion: $0.50/W is possible.
(3) Let's assume a fully-loaded hourly cost of an installation crew worker of $50 and that it will take a crew of 3 to install a 5kW system about 15 hours, that means labor costs add up to $2,250. Additionally, there is project management involved to manage the installation crews, communicate to customers and fulfill their unique wishes. Plus maintenance and leasing of trucks, etc. These costs can definitely be managed better and efficiencies can be achieved through innovation of faster installation techniques but the cost reduction curve is not going to be as steep as for hardware costs. Conclusion: $0.50/W is overly ambitious especially if you factor in wage inflation.
(4) Permitting costs are often underestimated. I wrote about permitting nightmares before here. I am hopeful that some cost reductions can be achieved through streamlining and standardizing permitting procedures in mature solar markets. Conclusion: $0.25/W is possible.
(5) The last cost bucket is a catch-all which includes sales, marketing and general business overhead. The costs of customer acquisition are particularly high in residential solar. Getting a homeowner's attention through advertising and other media is difficult, especially with a non-urgent product like solar electricity. Other general expenses are also high in the construction industry and difficult to reduce. Just look at Real Good Solar's financials. It's difficult to be profitable in this industry. Conclusion: $0.75/W is overly ambitious.
Overall, it becomes clear that there is a lot of potential for hardware costs to continue to decrease but I am doubtful about the potential to reach similar cost reductions for soft costs. As module and inverter prices decrease, the soft costs become a bigger and bigger percentage of total installed costs. Therefore, the cost reduction curve will flatten out. The good news is that a) even at $3.50/W, residential solar is cost competitive in large parts of the US, and b) soft costs means local jobs that cannot be exported to other countries.