Improvements to the indexing code could decrease that amount of lambdas to about 200 given some effort and hard work, so it seems like a hundred million files making use of the full 1,000 lambdas appears to be probable, but a lot more very likely it would be near to eighty million when carried out.