That particular sentence on the wiki is uncited, so trust it with extreme caution. That said, yes, anything is feasible with enough money, and generally you pay an engineer to determine what method will cost the minimum amount of money.

That's effectively what the Millenium Tower did - their engineers determined that friction piles to the sandy layer would be the cheapest acceptable option. Then, due to factors they may or may not have been able to predict, it turned out to not be an adequate solution.

If they had known beforehand, the next cheapest option would have been chosen - likely end-bearing piles to the bedrock, like they are now retrofitting. Floating raft foundations are truly a last resort due to their expense.

You are right - based on the numbers given at the start of the article, and a density of clay around 1700 Kg/M3, if one kept within the footprint of the tower, it seems one would have to excavate ~100 feet down to create a sufficient volume (less, of course, if you can broaden the chamber beyond that footprint, but I guess the scope for that is limited - and, on the other hand, the landfill layer may be less dense than clay.)

Big skyscrapers in London use a similar friction pile technique. E.g The Cheesegrater sits on friction piles which I believe are ø2.5m x 25m long.

There’s a suburb of Vancouver called Richmond that has been built on an old bog. There are building limits and other codes to deal with because of it but I heard recently that the ground in Richmond is actually becoming more stable of all the piles being driven into the ground from new condo developments and other infrastructure.

This is generally because of heave effects (clay rising up after loads being removed for basement excavation), and results in significant raft slabs (as much as 3m thick) and tension piles which can be horrendously expensive.