October 22, 2013 | 2
Today, the U.S. Department of Energy’s SunShot Incubator Program announced the award of $16M to 17 solar startups that are working to reduce the hard and soft costs of solar power.
Included in this latest round of SunShot awards was a company focused on the financing challenges of new solar installations. This solar startup – kWh analytics –wants to reduce the cost of financing solar installations by increasing investor confidence through access to data.
Reducing the cost to install solar panels remains a problem as available financing continues to be outstripped by demand. While total system costs have dropped dramatically, the growth in available capital has been sluggish. As a result, the solar industry is currently hobbled in its quest to become a more widely competitive energy resource.
Those who want to install solar face expenses including the cost of hardware, installation, permitting, metering, and financing along with any software that the owner adds for system management. Over the past five years, the price of solar hardware (panels) has dropped dramatically – from $4.50 per Watt in 2007 to $0.60 today.
As a result, the solar industry has increasingly focused on minimizing the soft costs of solar power. In 2011, the SunShot Initiative invested $13.6 million in seven projects that target reducing market barriers and non-hardware balance of system costs. These projects focus on streamlining project siting and permitting, decreasing headaches caused by common technical problems, and standardizing the installation process.
In the words of kWh Analytics CEO Richard Matsui, “understanding risk is an essential part of any investor’s job—but it’s been hard to do without data.” In the solar world, this means answering questions related panel quality, inverter reliability, and customer default rates so they can better place their bets.
A former McKinsey consultant and Forbes “30-under-30” in Energy, Matsui looks at solar in terms of how to quantify risks. And, he believes that sluggish investment volumes are largely the result of mis-perceived risk. When asked, Matsui and kWh CTO Nicholas Malaya provided the following discussion for Plugged In:
The Key to Unlocking Capital in the Solar Industry
By Richard Matsui and Nicholas Malaya, kWh Analytics
Today’s institutional investors are not comfortable investing widely in the solar power asset class because they don’t yet believe that they will realize the promised return on their investment. As a result, financing costs remain one of the largest remaining expenses for solar projects.
Solar investments are attractive to investors because the sun is of course reliable, and that the investments are also wholly uncorrelated with the rest of the financial markets—which is a highly unusual and valuable characteristic. But, investors still demand solar project IRRs in the double-digits, while other safe investments receive much lower returns.
That isn’t to say that the solar industry isn’t growing – of the 100 GW of solar currently installed around the globe, 2/3 of that was built in the last 2.5 years. But, this growth could be greatly accelerated by reducing project financing costs.
Two types of data are needed to help reduce financing costs and spur growth in the amount of capital available to solar projects.
First, data on the physical asset. That is, how much energy is generated in a specific location by a particular panel brand, inverter model, etc.
Second, data on the financial asset. Data related to the customer themselves to provide an indication of the chance that a person or company will pay back a loan – similar to a credit check before one is issued a credit card or mortgage.
As a result, we have aggregated the industry’s largest independent dataset of solar assets, pooling field performance data from 10,000+ PV systems representing nearly 1GW of generating capacity. That’s as much power as a nuclear power plant, but spread across thousands of sites with individualized weather patterns.
This dependency on weather patterns leads to intermittency in the energy generation from any particular solar installation. Accurately forecasting the energy production from fleets of PV systems is a substantial modeling challenge, requiring large volumes of historical performance data, and sophisticated statistical methods to characterize the many uncertainties present.
Controlling for factors such as weather, time of year, and the azimuth and inclination of a system, identical systems can be aggregated, resulting in distributions of expected system performance. This permits direct comparisons between the system components, such as the panel or inverter. These methods permit decision makers in industry to make better procurement decisions as well as providing more accurate system performance guarantees and predictions of maintenance costs and failure rates.
These predictions will grow more accurate as the industry installation base grows and the systems age. In time, subtle properties of system performance will be able to be inferred, such as the impact of dust accumulating on systems that have been cleaned recently. However, this will in turn present a new challenge, where the sheer volume of data generated from an enormous number of systems around the world threatens to overwhelm traditional databases and statistical programming packages. This ‘big data’ challenge will require distributed computing algorithms, such as Google’s MapReduce, in order to effectively process the data in a timely manner.
By assembling existing industry information into usable data products, we enable investors like banks an insurance companies to quantify solar project risks. And, through a data-driven understanding of the true risks of these projects, these investors can make smarter choices, deploy more capital, and reduce financing costs.
About the authors:
Richard Matsui – Co-founder & CEO of kWh analytics. He was a consultant with McKinsey & Company. He helped found their solar practice in 2007 and accumulated more solar experience than any other pre-partner consultant (9,000 globally). Richard graduated from Georgetown University magna cum laude and Phi Beta Kappa, and is also a Forbes “30 Under 30″ award recipient for Energy.
Nicholas Malaya – Co-founder & CTO. He is a research scientist at the University of Texas, with a focus in computer modeling on massively parallel machines. He has a ‘Thrust 2000′ fellowship and has run on several Top 10 supercomputers. He has been co-author on seven technical papers and given over twenty technical talks. Nick graduated from Georgetown University with a B.S. in Physics and Mathematics, where he was the recipient of the Treado Medal.