Loading
Of 3,679 open vaults, nobody has ever asked 2,645 of them what has left.
Every vault page in this industry answers what you might earn. This is the practical question underneath it, and the honest first thing to say is how much of it we cannot answer. An empty withdrawal history on a vault nobody read is our silence, not evidence that nobody got out.
The distinction this whole surface turns on. A vault we asked about and found no withdrawal in is a weak reading of that vault; a vault nobody asked about is not a reading at all, and the two look identical in any figure that reports zero for both. The gap between the two figures on this page is that distinction, counted: 1,034 have been asked and 1,021 came back with something, so 13 were asked and the answer was nothing. Those are the only vaults on this site whose empty withdrawal history means anything.
Six readings sit behind an exit assessment. Four of them are counted here, each over the same 3,679 vaults so they can be read against one another. The other two, the queue and the size you have in mind, exist only for a particular vault and a particular amount.
Has anybody actually left
1,021 vaults have had their movements read. Settled withdrawals are the only reading here about something that already happened, which is why it leads.
Is there anything free now
661 carry a withdrawable figure. On venues whose vaults lend into markets holding their own liquidity it is a FLOOR rather than a ceiling, and the assessment says so rather than reading it as the limit.
How much is lent out
623 are lending vaults whose book we can read, and 68 of those are at 95% utilisation or more. At that level a large exit depends on borrowers repaying rather than on cash in the vault.
Can somebody change the terms
426 have had their access control read from the chain. For the rest, who can move the money and how quickly is unread, which is not the same as nobody being able to.
“Can I get out” has no answer. “Can I get two million out of a vault with seventy-one million free and twenty-five thousand settled withdrawals behind it” has several, and they differ. vault_exit_confidence weighs those six readings against a size you name, and the answer is the WORST of them rather than an average: a queue in front of you is not offset by a long history, it happens before your turn. An unread component outranks a supporting one, because a gap we could not read must never be counted as a pass. No sentence it emits contains the word guarantee, and the register of what has actually left is free on vault_exits. You can run our cases against us first.