The Devil is in the Assumptions

Google just came up with a report on the potential of clean energy technology, which has received some fairly rapturous coverage in the environmental press. The key insights of the report, as follows:

• Energy innovation pays off big: We compared “business as usual” (BAU) to scenarios with breakthroughs in clean energy technologies. On top of those, we layered a series of possible clean energy policies (more details in the report). We found that by 2030, when compared to BAU, breakthroughs could help the U.S.:
• Grow GDP by over $155 billion/year ($244 billion in our Clean Policy scenario)
• Create over 1.1 million new full-time jobs/year (1.9 million with Clean Policy)
• Reduce household energy costs by over $942/year ($995 with Clean Policy)
• Reduce U.S. oil consumption by over 1.1 billion barrels/year
• Reduce U.S. total carbon emissions by 13% in 2030 (21% with Clean Policy)
• Speed matters and delay is costly: Our model found a mere five year delay (2010-2015) in accelerating technology innovation led to $2.3-3.2 trillion in unrealized GDP, an aggregate 1.2-1.4 million net unrealized jobs and 8-28 more gigatons of potential GHG emissions by 2050.
• Policy and innovation can enhance each other: Combining clean energy policies with technological breakthroughs increased the economic, security and pollution benefits for either innovation or policy alone. Take GHG emissions: the model showed that combining policy and innovation led to 59% GHG reductions by 2050 (vs. 2005 levels), while maintaining economic growth.

Well, hot damn. That's some good outcomes. All we need is a carbon price, some deployment policy, and a couple of scientific breakthroughs, and we can save the world and get rich at the same time. Well, I was feeling cynical, so I decided to look at exactly what breakthroughs we might need. Google was nice enough to publish it's data in Appendix C, so please turn down there, and look at solar PV. Google has the 2010 overnight capital costs--what it takes to build an electrical plant, at $4000/kW. The breakthrough scenario has that solar PV at $1000/kW in 2020. Batteries are another core technology for electric vehicles and grid-scale storage. Right now, Google has batteries at $500/kWh, and in their good scenario, $100 kWh in 2020. Other clean energy technologies see slightly smaller, but similar three or fourfold decrease in price in just a decade, along with major increases in reliability and lifespan. Now, I won't go out and say that those kinds of cost reductions are impossible, since prediction, especially about science and technology, is very hard. But in the case of solar PV, it would be about an improvement an order of magnitude better than what was seen in the past decade. In many cases it appears that we may be approaching limits imposed by the cost of raw materials: silicon, cobalt, lithium, steel, and rare earth metals. Without a better idea of what scientific breakthroughs are needed, or how those breakthroughs could be achieved, Google's report should be taken with a large grain of salt.