Binance-Peg BUSD Was Often Undercollateralized

DataFinnovation - ChainArgos - 4AC
Published in
5 min readJan 2, 2023


Binance runs their own bridge-esque setup which produces Binance-Peg tokens. These are tokens on a range of non-ETH blockchains backed by ERC20 tokens stored on ETH in a specific wallet. The process is described here. The wallet on ETH is here.

BUSD on BSC/BNB Smart Chain — the geth-fork chain rerun by Binance with many names — is not native Paxos-run BUSD per Binance’s docs. It is a Binance-Peg token. This means BUSD on BSC are backed by tokens in the wallet given above.

This wallet has had a lower balance of BUSD than the number circulating on BSC on many many days in the past. This means Binance-Peg BUSD was often undercollateralized.

Binance-Peg Wallet Flows

The backing wallet has only 22 BUSD flows. They look like:

If we add these up over time the balance is:

There is an odd flow of 1.854b out of the wallet on 17-Aug-2022 and then 1.851b back in on 23-Aug-2022. Other than that it just accumulates BUSD over time. More on that later.

BUSD Minting On BSC

Here we are going to look at BUSD during the large market cap growth in the first half of 2021. This is where BUSD growth went explosive per coingecko:

Here is BUSD minting on BSC during that time:

With a total balance as:

Now there is some weird action here in April 2021 where 1b are printed, then 800mm burned, and then an additional 2 prints of 100mm occur in quick succession.

So both sides are odd.

Comparing Them

These numbers vaguely track each other:

Note that the same peg wallet backs BSC, BNB Beacon Chain — the Cosmos side of the BNB chain pair — Avalanche, Polygon and TRON. The balances on the last 3 chains are tiny. But BNB Beacon Chain has a large amount. So we might expect to find the red line well over the green one.

But the green line over the red one means there are more BUSD circulating on BSC than exist in the ETH backing wallet even ignoring all the other chains. Here is the difference for this time range:

Negative numbers mean Binance-Peg BUSD is undercollateralized. And we can see this happened a lot. And not by small amounts.


First, this was a while ago. The Binance-Peg process is sufficiently backed by now and has been for some time.

Second, even at that time Binance wallets held most BUSD so it is possible these were backed in some other way. It is also possible those other wallets contained client deposits and shouldn’t be counted as backing. We do not know.

What we do know is that they’ve laid out a specific process for Binance-Peg and that process was not followed sufficiently-well to achieve full collateralization. Proof of this problem exists on-chain.

Third, there are odd flows associated with this thing getting regularized. Look at this pair of flows. Why would they withdraw and then redeposit 1.85b BUSD on ETH? And check out this pair on BSC. Why would they mint and then burn 1b->800mm BUSD there?

This is not proof that unbacked stablecoins were printed and then, later, backed with client deposits. But it is proof that unbacked stablecoins were printed and subsequently got backed by some source of funds inconsistent with the stated process. And it is proof the flows through the backing wallets do not make much sense with respect to the process Binance was supposed to follow.

BUSD on BSC has a history of being significantly unbacked.

BNB Price Impact

Now BNB is the native token — it is the thing you use to pay gas — on the geth-fork Binance Smart Chain. Let’s plot the price of BNB during the period of aggressive unbacked printing of BUSD on that chain:

The run up in price begins exactly when the dramatic unbacked printing starts. Simialrly the price stops rising, and then retraces a bit, when the printing stops.

This does not prove anything of course. Doubly so as the BUSD are eventually backed by ERC20 tokens in the peg wallet. But it is suspicious and indicates that pumping of BNB with unbacked BUSD might have occured. At best — at absolute best — Binance was careless during this time as the tick marks on the horizontal axis are almost 7ish weeks apart.

Further this activity occurred nearly 2 years ago in a major stablecoin, branded by the largest exchange, on a major chain. There is surely a lot still to be discovered.