I chose a random sample. In scientific studies we always do that. The assumption is a randomly chosen sample should represent the population! This is fundamental of statistics.

Here, I have chosen another random sample and made a plot.

Do you see any difference? I don't. I can look at a bigger sample, I can look at half of the population and I can look at the entire population. I will find the same answer.

~57% of the accounts hold less than 1 hive or 20 Leo worth of assets

It matters not if all the Leo users cumulatively hold 10M hive or 20M hive. The above hypothesis is solid so far. Even if that number is half that, say 28%, that is still dismal!