By Nitesh
- A Bitcoin sidechain is a separate blockchain that issues a token backed by Bitcoin locked on the main chain.
- You do not hold Bitcoin on a sidechain. You hold a claim against whoever controls the locked reserve.
- Most Bitcoin sidechains are run by a federation . a fixed set of signers, typically between 11 and 15 entities.
- The main chain will never reorganise to fix a sidechain failure, which makes losses on that side final.
A Bitcoin sidechain exists because Bitcoin deliberately does very little. No smart contracts of consequence, ten-minute blocks, limited throughput. Those constraints are the security model, not oversights.
A sidechain is the workaround: a separate chain with different rules, holding a token that is supposed to be worth one Bitcoin.
What is a Bitcoin sidechain, in plain English?
A Bitcoin sidechain is a second, independent blockchain with its own blocks, its own validators or signers, and its own rules about speed and features.
Bitcoin cannot move onto it. Instead, real Bitcoin is locked on the main chain, and the sidechain issues an equivalent token against that lock. Move Bitcoin in through a peg-in, take it out through a peg-out.
The sidechain can then do things Bitcoin will not: confirm in seconds, run confidential transactions, issue other assets. The main chain never learns any of it happened.
Who holds the Bitcoin while you use the sidechain?
Not you. This is the part most people get wrong.
Liquid’s own statement says the SideSwap key was not compromised, nor were any others.
That sentence, published during this month’s incident, is the clearest statement of where a Bitcoin sidechain’s risk actually sits. The keys held. Something else did not.
The locked Bitcoin sits in a multi-signature wallet controlled by a federation: a fixed group of named entities running hardware security modules. Liquid, the best-known Bitcoin sidechain, uses fifteen such signers, and a quorum of eleven is enough to authorise a peg-out.
What you hold on the sidechain is an IOU. It is worth one Bitcoin because the federation will honour it, and for no other reason. This is exactly the distinction we set out in custodial versus non-custodial storage, applied at the level of a whole chain rather than a wallet.
How is that different from a Layer 2?
A real Layer 2 inherits the main chain’s security. A Bitcoin sidechain does not.
On the Lightning Network, funds are secured by Bitcoin scripts, and you can always force settlement back to the main chain unilaterally. Nobody’s permission is needed. On a sidechain, exit requires the federation to sign. If they refuse, or if their software misbehaves, you have no unilateral route home.
That single difference is why the two should never be described interchangeably, even though both are marketed as scaling Bitcoin. If you want the underlying mechanics, our explainer on how crypto bridges work covers the same lock-and-mint pattern.
How does a Bitcoin sidechain actually fail?
A Bitcoin sidechain rarely fails through stolen keys. Usually by a bug that makes an invalid request look valid.
If the code that issues sidechain tokens can be tricked into minting tokens with nothing locked behind them, the peg-out process will faithfully pay real Bitcoin against those fake claims. Every signature verifies. The federation is not compromised; it is answering a question that was framed dishonestly.
That is precisely what happened this month. As we reported, roughly 3,998 BTC left Liquid’s federation wallet in a single transaction after tokens were created through a flaw in the Elements software. Eleven or more of fifteen signers approved it, and Blockstream confirmed no key had been compromised.
Why can’t the main chain undo it?
Because Bitcoin does not know the sidechain exists.
Some chains can reverse a hostile transaction if enough validators agree. We covered one: Cronos discarded 10,961 blocks to undo a $75 million exploit. That is possible with a hundred validators who can be convened overnight.
Bitcoin will not do this for anyone. A peg-out that lands on the main chain is final, and recovery depends entirely on persuading whoever holds the funds to return them. In Liquid’s case that meant negotiating, and about 85% came back while roughly 598 BTC did not.
What happens step by step when you peg in?
Worth walking through, because the mechanics explain where the trust sits.
You send Bitcoin to an address the federation controls. That address is a multi-signature output requiring a quorum of the federation’s keys to spend. Your Bitcoin is now locked, in the sense that no single party can move it.
The federation’s nodes observe that deposit. After a confirmation threshold, typically around 102 blocks on Liquid, they collectively authorise the sidechain to issue an equivalent amount of the sidechain token to an address you control on that chain.
The delay is deliberate. A shorter wait would expose the federation to a Bitcoin reorganisation, where a deposit that appeared confirmed gets orphaned after tokens have already been issued. Roughly seventeen hours of waiting is the cost of eliminating that risk.
Peg-out reverses it, and this is where it gets asymmetric. You send the sidechain token to a burn address, the federation observes it, and a quorum signs a Bitcoin transaction releasing the locked coins. If the quorum does not sign, nothing happens. There is no timeout, no fallback, and no unilateral escape hatch.
What is a federation, exactly?
A fixed set of named organisations, each running a hardware security module, that jointly control both the locked Bitcoin and the sidechain’s block production.
On Liquid these are called functionaries, and there are fifteen. Signing a peg-out requires eleven or more. Block production rotates among them, which is why the chain produces blocks every minute rather than every ten.
The membership is public and the participants are mostly exchanges and infrastructure businesses with reputational and commercial stakes in the chain working. That is genuinely a security property. It is also the ceiling on how decentralised the arrangement can be: fifteen known entities, in identifiable jurisdictions, subject to identifiable legal process.
Compare that with Bitcoin itself, where no list of participants exists and no jurisdiction can compel the network. The comparison is not a criticism of sidechains; it is the trade being made in exchange for speed and features.
What can a sidechain do that Bitcoin cannot?
Four things that matter commercially.
Fast settlement. One-minute blocks with deterministic finality after two blocks, against Bitcoin’s ten-minute probabilistic confirmations. For an exchange moving inventory between venues, that difference is operational rather than cosmetic.
Confidential transactions. Liquid conceals both the amount and the asset type of a transfer while remaining publicly verifiable. Bitcoin exposes every amount to anyone. For an OTC desk moving size, visible amounts are a real disadvantage.
Issued assets. A sidechain can host tokens beyond its pegged Bitcoin: stablecoins, security tokens, tokenised instruments. Bitcoin’s base layer cannot practically do this.
Lower and more predictable fees. A sidechain’s block space is not being auctioned against global demand, so a transfer costs a fraction of a cent regardless of what the main chain is doing. Our explainer on how gas fees work covers why main-chain costs spike unpredictably.
How do the trust models compare in practice?
Three arrangements, three failure modes, and it is worth being precise about which is which.
Bitcoin main chain. You hold the key, you control the coins, nobody can freeze or reverse them. The failure mode is your own error, covered in what a private key is.
Lightning Network. Funds sit in a two-of-two channel secured by Bitcoin script. Your counterparty cannot steal them, and if they disappear you close the channel unilaterally and settle on the main chain. Bitcoin’s own rules enforce this. The failure mode is liquidity and routing, not custody.
Sidechain. Funds sit with a federation. Exit requires their cooperation. The failure mode is the federation, whether through collusion, legal compulsion, or a software bug that makes a fraudulent request look legitimate.
Notice that only the third has a failure mode where you did nothing wrong and the counterparty was not malicious. That is the specific risk a sidechain adds.
What did the September 2026 failure actually reveal?
That the federation model protects against the wrong threat.
Everyone assumed the danger was key compromise, so the design put fifteen keys in fifteen hardware modules across fifteen organisations. That defence worked perfectly. Blockstream confirmed no key was compromised.
The failure came one layer earlier. A flaw in the Elements software allowed sidechain tokens to be created without corresponding Bitcoin being locked. The peg-out process then did exactly what it was built to do: verify that the request was properly authorised, confirm it was, and release real Bitcoin.
Eleven or more functionaries signed a valid transaction paying out against invalid tokens. The reserve fell from 4,200 BTC to 207.275, roughly 95% of the collateral behind every outstanding sidechain token.
The instructive part is that no amount of additional signers would have prevented it. Twenty-five functionaries would have produced the same result, because the question they were asked was correctly answered. The vulnerability was in the issuance logic, and signature thresholds do not police issuance.
Why is a peg-out final when other chains can roll back?
Because the two failures happen on different ledgers.
When a chain controls its own history, a large enough quorum of its validators can agree to discard blocks and rebuild from an earlier state. We covered exactly that when Cronos discarded 10,961 blocks to reverse a $75 million exploit, erasing almost two hours of everyone’s transactions in the process.
A sidechain peg-out lands on Bitcoin. Bitcoin’s validator set is enormous, anonymous and has never reorganised to remedy anyone’s loss. There is no committee to petition. Once the transaction confirms, the Bitcoin is gone in the only sense that matters.
Which is why the Liquid incident resolved through negotiation rather than engineering. Roughly 3,400 BTC came back after on-chain talks; about 598.5 BTC did not. We covered the outcome in why the hackers kept 598.5 BTC.
What happens to the token if the reserve is short?
Nobody has answered this cleanly, and that absence is itself the important fact.
If outstanding sidechain tokens exceed locked Bitcoin, holders are creditors of a reserve with a hole in it. Three outcomes are possible. The operator absorbs the shortfall from its own balance sheet. Holders share it pro rata, and the token trades below par. Or peg-out is suspended until the gap is closed by other means.
Which of these applies is generally not documented in advance. There is no deposit insurance, no resolution regime, no published loss-allocation waterfall. You discover the answer after the event, from an operator under pressure.
The market signal to watch is simple: whether the token trades at one Bitcoin when peg services reopen. A discount is the market pricing the shortfall, and it is visible within hours.
How should you size a position on a sidechain?
Treat it as a credit exposure to the federation, not as a Bitcoin holding, and size it accordingly.
A practical rule used by trading desks: hold on a sidechain only what you would be willing to leave with a single well-run counterparty overnight, and only for as long as the transaction requires. That is typically hours, not months.
The specific questions worth answering before you use one at all:
How many signers, and what is the quorum? Then remember that this number protects against key theft and nothing else.
Is the codebase audited, and by whom? The Elements failure was in issuance logic, so ask specifically whether minting and peg validation were in scope.
How long has the peg held real value? Years of holding nine figures under continuous attack is the only meaningful audit.
What is the documented loss-allocation policy? If there is none, assume you bear it.
Can you exit without permission? On a sidechain the answer is no. If that answer matters for the funds in question, use the main chain or Lightning instead.
Should you use one?
For moving value quickly between exchanges a Bitcoin sidechain is genuinely useful, and that is what sidechains are genuinely good at. Settlement in seconds for a fraction of a main-chain fee is a real service.
For holding, no. Ask three questions before leaving anything on one:
How many signers, and who are they? Eleven of fifteen sounds robust until you realise it protects against key theft and not against a minting bug.
How long has the peg held real value? Time under attack is the only audit that counts.
What happens if the reserve comes up short? Whether the operator absorbs the loss or holders share it is usually undocumented, and you find out afterwards.
The rule is the same one that applies to every wrapped asset. A sidechain token is a credit position, and it should be sized like one. Bitcoin’s own design is covered in our guide to how Bitcoin works, and the case for holding keys yourself in what a private key is. Technical documentation for the sidechain model is published by the Elements project, and Bitcoin’s base-layer rules by bitcoin.org.
What's your reaction?
