The decisive step in the Binance hack may have been a swap, not a theft alone. Before the freezes, about $786M in stablecoins was converted into ETH and TRX, two assets that no central issuer can freeze.

USDT and USDC are different in kind. Tether can freeze USDT and Circle can freeze USDC, because each token has a central issuer. ETH and TRX have no issuer with that power, so the conversion removed the converted funds from the freeze mechanism entirely.

Issuers later froze substantial sums: about $1.567B in USDT from Tether and about $823M in USDC from Circle. Elliptic estimates the total frozen at about $2.84B, while Binance says about $3B was frozen.

Those totals need to be read beside the overall hack, estimated at about $6.3B, and beside Elliptic's estimate that the attacker still controls about $3.33B in assets that cannot be frozen by a centralised token issuer.

The conversion helps explain the size of that remaining figure. It was not simply that some tokens escaped notice; a large sum was deliberately transformed into a form that issuers cannot seize.

The Aave figures point in the same direction. Per LlamaRisk, about $59.9M in stolen USDT was deposited as collateral there and about $43.8M worth of ETH was borrowed against it, another route from a freezable token into ETH.

The lesson is structural. In a decentralised ecosystem, centralised freezes can be large and fast, yet they stop at the edge of issuer-controlled tokens, and the ETH and TRX conversion shows how much value can already sit beyond that edge.