There are many inactive wallets out there because the private keys have been lost forever. Grouping these in with "control" isn't really appropriate since no human controls them.
Quit a bit of the early coins were lost. Since they had no value people just threw them away when installing the new client software. It's not like there was an exchange or anyone willing to pay real dollars or any serious expectation that anyone would. So a study of the bitcoin blockchain that just looks at wallet addresses would be kinda flawed. They were very easy to mine with a desktop CPU and you'd get tons of the worthless things in the early days.
You can conclude from this: 1) There are many fewer bitcoin whales than the numbers make it seem, and 2) The total supply of tradable bitcoins is significantly smaller than the blockchain shows in wallets.