In early September 2026, two Bitcoin-linked protocols were hit within five days of each other. Liquid Network, a Bitcoin sidechain operated under Blockstream’s stewardship, saw $320 million in BTC — roughly 4,000 of the 4,200 BTC in its reserve — withdrawn by attackers who had deposited nothing. Days later, Symbiosis, a cross-chain liquidity protocol, halted its Bitcoin bridge after an attacker minted an absurd quantity of unbacked syBTC and cashed out for about $336,000.

Neither attack involved phishing. No victim clicked a link, connected a wallet to a drainer, or signed a malicious approval. The attackers aimed somewhere else entirely: the protocol’s own logic for verifying that a deposit happened.

This article dissects both incidents, extracts the shared pattern, and walks through what — if anything — an ordinary holder can actually do about it. If you want the broader taxonomy of bridge attacks, our earlier pieces on bridge attacks and bridge security assessment cover the full landscape; here we zoom in on the single technique that just took down two more protocols.

Case 1: Liquid Network — forging the deposit receipt on a sidechain

Liquid Network is a sidechain to Bitcoin. Users lock BTC with the network’s federation and receive L-BTC in return — tokens that Chainalysis, in its official post-mortem, described as “depository-receipt-like assets” fully backed by BTC under normal conditions. When you spend or move L-BTC, you are passing around a claim on real Bitcoin sitting in the reserve. The same model underpins most wrapped assets in crypto, including WBTC.

The vulnerability, per Chainalysis’s reconstruction, lived in Liquid’s transaction-validation software. In simple terms, the system could be tricked into accepting L-BTC that was never backed by a real deposit. The attackers minted unbacked L-BTC and redeemed it against the reserve for real BTC — essentially exchanging an IOU they had fabricated for other people’s Bitcoin.

What happened next was unusual. The attackers contacted Blockstream, described themselves as white hats, and said they would return most of the funds once the bug was patched. Blockstream deployed a fix; by Wednesday, September 9, 85% of the withdrawn BTC had come back. Roughly $47 million remained with the attackers, with negotiations continuing, according to a statement from Liquid Network. Per public reporting, Blockstream’s co-founder Adam Back said the L-BTC peg would be fully covered.

The important detail for holders: L-BTC did not collapse because the reserve was made whole. Had the attackers not returned anything, every L-BTC holder would have been holding claims against a reserve with a hole in it — the mathematics of a fractional backing nobody agreed to.

Case 2: Symbiosis — a mint of 2^62 units, a cash-out of $336,000

Five days later, at roughly 04:28 UTC on September 11, an attacker sent an abnormal message to Symbiosis’s BridgeV2 contract on BNB Chain. The contract accepted it. The result was a mint of approximately 2^62 raw units of syBTC — Symbiosis’s synthetic Bitcoin token — an amount whose notional face value ran into tens of billions of dollars. More unbacked syBTC was minted than Bitcoin that will ever exist.

The cash-out was far more modest: about 4.39 WBTC, roughly $336,000 in realized value. DeFiLlama’s hack tracker classifies the incident under “Bridge & Cross-Chain” with the technique recorded as an unbacked cross-chain mint. Symbiosis suspended its BTC routes and kept other routes running.

The gap between the $46-billion-notional mint and the $336,000 cash-out is worth sitting with. The attacker’s constraint was not the bug — the bug let them print effectively unlimited synthetic Bitcoin. The constraint was the size of the pools they could dump into. That is not a defense. It is luck, and it does not scale down risk: a smaller reserve simply means smaller losses per incident, not fewer incidents.

The shared pattern: attack the verifier, not the victim

Strip away the specifics and both incidents are the same three-step play:

  1. Forge the evidence of a deposit. Craft a message or transaction that the protocol’s verification logic will accept as proof that real assets arrived. No assets actually arrived.
  2. Mint the pegged asset. The protocol, believing its own records, mints the wrapped or synthetic token against the fake deposit.
  3. Redeem for the real reserve. Exchange the freshly minted token for real assets held in the reserve pool — the BTC that legitimate users deposited.

Compare this to the attacks most security education prepares you for. A phishing site or a wallet drainer needs a human to fail: to click, to connect, to sign. Every guide we publish about reading signatures, isolating claim wallets, and verifying URLs targets that failure point. An unbacked mint attack has no such failure point to exploit. The only human who has to fail is a protocol engineer, and they failed months earlier, in code that was deployed and running.

This is why the unbacked mint pattern deserves its own name rather than being filed under “bridge hacks.” The attack surface is the verification logic itself — the code whose entire job is to distinguish real deposits from fake ones.

Who the victim actually is

In a phishing attack, the victim is obvious: the person who signed. In an unbacked mint attack, the victim is collective — every holder of the pegged asset, at once.

The reserve backing a wrapped asset is shared collateral. When an attacker redeems fake receipts for real reserve assets, they dilute the backing of every token in circulation. If the drain is large enough and not made whole, the token depegs — not because of market panic, but because the arithmetic of backing no longer works. You did nothing wrong, you signed nothing, you clicked nothing, and your L-BTC or syBTC is now a claim on a smaller pile of real Bitcoin than the issuance promised.

Liquid Network’s outcome — 85% returned, peg publicly committed to full backing — is the good ending. Symbiosis’s $336,000 is small enough to be absorbed. But “the attacker gave it back” and “the number was small” are not security properties. They are outcomes.

What a holder can realistically do

Honest answer first: you cannot audit a protocol’s transaction-validation software yourself, and neither can most security firms — this class of bug survives audits regularly. The defenses available to a non-engineer are about exposure, not prevention.

  • Prefer native assets when you don’t need wrapped ones. Every day you hold a wrapped or synthetic asset is a day you are a silent creditor of that protocol’s verification logic. Bridge in, transact, bridge out.
  • Treat holding time as risk. The exposure to an unbacked mint is not the transaction — it is the balance sitting in the wrapped form between transactions.
  • Watch reserve health, not just TVL. A proof of reserves attestation tells you reserves existed at audit time. It says nothing about whether the minting logic is correct. DeFiLlama’s stablecoin and wrapped-asset pages show backing ratios; a pegged token trading persistently below par is the market pricing in a backing question.
  • Take emergency halts seriously. When a protocol suspends a route — as Symbiosis did — the price of its synthetic asset on secondary markets is an immediate referendum on whether the market believes the backing is intact.

Limitations of this analysis

We reconstruct both incidents from public reporting: Chainalysis’s post-mortem for Liquid Network, and initial reporting plus on-chain alerts (Blockaid flagged the 2^62 syBTC mint) for Symbiosis. Neither protocol has published a full technical root-cause analysis of the exact code paths at the time of writing; the Symbiosis post-mortem in particular is still pending. The “white hat” framing of the Liquid actors is their own characterization — the $47 million still outstanding as of September 9 sits uncomfortably next to it. Details may shift as disclosures land; the pattern, however, is already fully legible.

Key takeaways

  • Unbacked mint attacks bypass the user entirely: no phishing, no signatures, no stolen keys — just a forged deposit record and a protocol that believed it.
  • Liquid Network (Sept 6, $320M withdrawn, 85% returned) and Symbiosis (Sept 11, ~$336K realized) are the same technique at different scales, five days apart.
  • The victim is every holder of the pegged asset, whose shared backing was drained.
  • User-side security — hardware wallets, signature discipline, phishing awareness — offers zero protection here. Exposure management is the only lever a holder controls.
  • Proof of reserves does not cover this failure mode; it audits reserves, not mint logic.

For the wider family of cross-chain failure modes — validator key compromise, signature threshold manipulation, and where fake deposits sit among them — read our bridge attacks breakdown. For how pegs break once backing comes into question, see our piece on stablecoin depegging.

Frequently Asked Questions

What is an unbacked mint attack?

An attack where the attacker never deposits the real asset but convinces a protocol's validation logic that a deposit happened. The protocol then mints a wrapped or pegged asset out of thin air, which the attacker redeems for real assets from the reserve pool. It is an attack on the protocol's verification code, not on any user's wallet.

Did Liquid Network users lose their Bitcoin?

As of September 9, 2026, the purported white-hat group had returned 85% of the roughly 4,000 BTC withdrawn (about $47 million remained outstanding), Blockstream had patched the vulnerability, and the network restarted. According to public reporting, Blockstream's co-founder stated the L-BTC peg would be fully backed. Holders did not lose funds to date, but only because the reserve was returned — the reserve itself, which backs every L-BTC holder, was the thing that got drained.

How is an unbacked mint attack different from phishing or a wallet drainer?

Phishing and wallet drainers attack the user: they need you to click, sign, or leak a seed phrase. An unbacked mint attack never touches the user. The attacker sends one crafted transaction or message to the protocol itself, exploiting a bug in how the protocol verifies deposits. No amount of user education, hardware wallets, or signature-reading discipline can stop it — the vulnerability is in the protocol's code.

Can proof of reserves prevent unbacked mint attacks?

No. Proof of reserves verifies that reserves exist at a point in time. An unbacked mint exploits the minting logic itself — the mint happens with the reserve untouched, and the drain follows immediately when the fake receipt is redeemed. A protocol can pass every reserve audit and still be drained in minutes if its validation code accepts a forged deposit message.