The news has been grim for Lehman Brothers today, but one part of the giant investment bank that's doing yeoman work is its solar energy equity research group, led by analyst Vishal Shah. His daily email newsletter and periodic reports are some of the most informative and cogent in the space.
Earlier this week, he sent out a detailed industry overview/solar demand analysis, including major takeaways from the recent EU PVSEC conference and exhibition in Valencia, Spain. I asked Vishal if it would be OK to excerpt some of the report, and he said to "feel free" to do so.
I've extracted one subsection of particular interest for the semiconductor community--and those semi refugees now working in solar. In my blog at sister site PV-Tech, look for different excerpts.
On why the solar industry is not similar to the DRAM industry, the Lehman report says the following (with minor editing):
The solar industry is often compared with the DRAM industry—common comparisons include the commodity nature of both DRAM and solar industry, the relatively low technology barriers to entry, and low-cost manufacturing as the key differentiating factor among companies.
We believe the solar industry is different from DRAM for the following reasons:
- Solar industry revenue pool is growing whereas DRAM industry revenue pool is shrinking.
- Fixed costs for DRAM manufacturers are high whereas fixed costs for solar companies are low.
- DRAM industry has no pricing floor, solar industry has pricing floor.
- Capital intensity of DRAM industry is twice that of solar industry.
- No Moore’s Law within solar industry–where is the need for replacement capex every few years?
But the report does note that "within the solar industry value chain, the polysilicon industry perhaps represents the closest [thing] to the DRAM industry." Here's why Vishal and his team think so:
In our view, there is a high level of similarity between polysilicon industry and DRAM industry as both industries are characterized by high capital intensity levels. We believe due to the relatively high fixed cost structure, polysilicon manufacturers are likely to continue to run factories at nearly full utilization rates as long as polysilicon price is greater than marginal cost. We believe this competitive dynamic is likely to result in significant polysilicon price declines, just as we have observed in the DRAM industry in the prior cycles.