9% is an absurd failure rate for solid state electronics. Particularly considering the profit margins. I assume it's related to the power densities involved. Would you happen to recall the source?