Multichain USDC: the bridge died, not the dollar

Multichain (formerly Anyswap) was a cross-chain router. To put USDC on Fantom, you locked native USDC in Multichain's wallet on a source chain and the bridge minted a wrapped representation — USDC.m — on Fantom. The wrapper was supposed to be redeemable one-for-one for the locked deposit, and as long as that redemption was credible, USDC.m traded at a dollar and circulated through Fantom DeFi as if it were the real thing.

On July 6, 2023 the locked deposits left. Around $126–130M of assets drained from Multichain's Fantom, Moonriver, and Dogechain bridge contracts in a single abnormal burst, with nearly $120M coming off the Fantom bridge — wBTC, wETH, USDT, and roughly $62M of USDC. Security analysts found no smart-contract bug; the signature pointed to a compromise of the bridge's multi-party-computation keys, the credentials that authorized moving the locked funds.

Because USDC.m existed only as a claim on those now-empty wallets, it repriced to what an unbacked claim is worth. It fell to around 22% of face, near $0.22, and dragged Fantom's DeFi TVL from roughly $364M down toward $69M. Circle froze about $63M of the underlying USDC on Ethereum, but that protected the real dollars — not the holders of the wrapper. Multichain shut down within two weeks, and USDC.m never recovered.

The crucial fact is that USDC.m and USDC were never the same asset, even though they shared a ticker and a price chart. Native USDC is a fiat-backed token whose dollar reserve sits with Circle. USDC.m was a derivative of that token: a Fantom-chain claim whose backing was a pile of real USDC locked inside a bridge wallet. The wrapper inherited none of Circle's reserve guarantees — it inherited the bridge's custody risk.

That distinction is invisible in calm markets and decisive in a crisis. When the bridge's deposits were swept out, the native USDC on Ethereum stayed fully backed and held its peg without a flicker, while the Fantom claim on those deposits cratered. The token that wore the stablecoin's clothes failed; the dollar underneath it did not. Anyone who read USDC.m's stable price as evidence of USDC's reserve quality was reading the wrong layer.

This is the missing failure mode that a fiat-cash lens alone does not capture. A wrapped or bridged stablecoin carries the issuer's reserve risk plus the bridge's counterparty and infrastructure risk, and in this case the second risk was the one that bit. Pharos treats bridge-wrapped representations as distinct exposures rather than as interchangeable with their native counterparts, which is why USDC.m sits in the cemetery while USDC trades on.

Multichain marketed itself as decentralized, secured by multi-party computation so that no single party could move the locked funds. The post-mortem showed that was a fiction. The MPC node servers all ran under founder Zhaojun's personal cloud account, and no other team member could log in. The cryptographic ceremony was distributed; the servers running it were not.

So when Zhaojun was detained around May 21, 2023 and his computers, phones, hardware wallets, and seed phrases were confiscated, the bridge lost its only operator. The team could not maintain the infrastructure, could not rotate keys, and could not protect the deposits. Whoever ended up with the confiscated credentials — the outflows have been attributed to an insider rug pull as much as an external hack — held unilateral control over hundreds of millions of dollars of user funds.

The lesson is that decentralization is an operational property, not a marketing claim. A bridge whose continued solvency depends on one person's freedom and one person's cloud login is a custodial service with extra steps, and its wrapped tokens carry exactly that custodial risk. The label said trustless; the architecture said trust one man.

Through the entire episode, real USDC held its dollar peg. Circle's reserve of cash and short-dated Treasuries was never involved — the assets that disappeared were USDC deposits that holders had voluntarily locked into a third party's bridge. The failure was downstream of Circle entirely, in infrastructure Circle neither built nor controlled.

Circle did intervene, but in the one way it could: it used the USDC contract's freeze function to blacklist about $63M of the stolen underlying USDC on Ethereum, hampering the attacker's cash-out. That action defended the integrity of native USDC and clawed back real dollars for eventual recovery. It did nothing for USDC.m holders on Fantom, who held claims on a wallet that no longer contained their backing.

The clean contrast is the whole point of the case. One asset is a fiat-backed stablecoin with a reserve and a freeze switch; the other was a bridge IOU with neither. They diverged by roughly 78 cents on the same day because they were never the same risk — and the price chart, which had shown them in lockstep for months, had been hiding that the entire time.

A wrapped stablecoin is not the stablecoin it names. It is a claim on a bridge, and it carries the bridge's custody, key-management, and operator risk on top of whatever the native issuer brings. Track the wrapper as its own exposure: who holds the locked backing, on what infrastructure, and what happens to your claim if that infrastructure goes dark.

"Decentralized" is a property you verify, not a label you accept. If a single account, key holder, or individual can move the collateral or take the servers offline, the protocol is custodial regardless of the cryptography wrapped around it. Multichain's MPC was real and its decentralization was not, and the gap between the two is where the money went.

Counterparty failure can depeg a wrapper while the underlying asset is perfectly healthy. The native dollar never moved; the bridge that held it did. When evaluating a depeg, separate the question "did the backing fail?" from "did access to the backing fail?" — for bridged assets, the second is usually the one that kills you.