Key Takeaways
- What it is: The movement of assets between chains, which is three different mechanisms wearing one word.
- How it is built: From an escrow balance under lock-and-mint, from paired burn and mint events under burn-and-mint, and from the protocol's own intent records under a liquidity network.
- Core expression: For lock-and-mint only, escrow balance = sum of representation supply across destination chains, plus retained fees, plus in-flight legs.
- Best use: Verifying that outstanding representations are backed, and measuring how much of an asset is present on a given chain.
- Main limitation: Aggregate bridge total value locked adds three incompatible quantities, and gross volume double counts round trips and operator rebalancing.
- Practical rule: Identify the design before reading any number. If you cannot say which of the three you are looking at, you cannot say what the number means.
Who This Guide Is For
Read this page if you have compared two bridges by total value locked, or read a chain's bridge inflow as evidence of capital rotation. Both comparisons assume the underlying mechanism is the same across the things being compared, and it usually is not.
Escrow reconciliation is the only continuously verifiable safety invariant a bridge exposes. What it implies for tradeable depth on a destination chain is taken up separately on liquidity fragmentation.
Educational content. Not individualized financial advice.
What Does Crypto Bridge Flows Measure?
It measures the movement of value between chains, but it measures three different things depending on the design, because no chain can verify another chain's state directly.
The three designs
- Lock-and-mint. The asset is locked in an escrow contract on the source chain and a representation is minted on the destination. The asset never leaves the source chain; a claim on it travels instead. Return trips burn the representation and release the escrow.
- Burn-and-mint. The token is destroyed on the source chain and an equal amount created on the destination, using mint authority the protocol holds on both sides. Total supply across all chains is constant, and no escrow exists anywhere.
- Liquidity network. The user deposits into a contract on the source chain and is paid out of a pre-funded pool on the destination chain, generally within seconds. The asset does not cross. The operator rebalances the two pools later, in bulk, and possibly through an entirely different route.
Plain-language definition
Bridge flow is how much of an asset moved from one chain to another, except that under one of the three designs nothing moved and somebody paid you locally instead.
Technical definition
Under lock-and-mint the measurable object is a stock: the escrow contract's balance at a block, readable exactly like any other contract account balance. Under burn-and-mint the measurable object is a matched pair of events on two ledgers, joined by a message identifier the protocol carries between them. Under a liquidity network the measurable object exists only in the protocol's own intent records, because the two on-chain legs share no asset and no counterparty.
How Is Crypto Bridge Flows Constructed?
| Design | Is there a readable escrow? | How volume is measured | What its total value locked actually is |
|---|---|---|---|
| Lock-and-mint | Yes, one address set on the source chain | Lock and unlock events | Cumulative net asset bridged out of the source chain |
| Burn-and-mint | No, nothing is escrowed | Burn and mint events paired by message identifier | Structurally zero, and often omitted or reported as blank |
| Liquidity network | A pool balance exists but holds no bridged asset | Fulfilled intents recorded by the protocol | Operator working capital, funded by liquidity providers |
- Classify the design before touching any number, by checking whether an escrow address holds the asset and who controls mint authority on the destination representation.
- For lock-and-mint, enumerate every escrow address and every destination representation, then read both at a fixed block per chain.
- For burn-and-mint, collect burn events on each chain and mint events on each other chain, and join them on the protocol's message nonce rather than on amount and timing, which produces false pairs.
- For a liquidity network, accept that chain data alone cannot reconstruct the transfer, and use the protocol's records while stating that dependency.
- Record the block per chain, since a cross-chain snapshot taken at one wall-clock time sits at different heights on each chain.
Formula and Measurement Logic
Lock-and-mint invariant: escrow balance = sum of representation supply across destination chains + retained fees + in-flight legs
The identity gives a backing ratio, defined as escrow balance divided by total representation supply, which should sit at or slightly above one. It is checkable at every block by anyone, without cooperation from the operator.
Burn-and-mint has a weaker conservation law: the sum of the asset's supply across all chains is constant, so a burn on one chain must be matched by a mint on another. That statement is true and almost useless as a monitor, because it holds only after every in-flight message has landed, and there is no moment at which none are in flight.
A liquidity network has no conservation law readable on chain at all. Its pool evolves as pool balance carried forward, plus inbound fills, minus outbound fills, plus rebalancing that may arrive through an unrelated route or off chain entirely.
Two directional quantities apply to every design. Gross volume is the sum of both directions, which is what most published figures report. Net bridged position is source-to-destination minus destination-to-source, and under lock-and-mint that net equals the change in escrow balance exactly, which makes the two an independent cross-check on each other.
How Should Crypto Bridge Flows Be Interpreted?
Bridge volume answers how much traffic a route carried. Bridge locked value answers how much asset sits in an escrow. Neither answers the question people usually ask.
The question people mean
If you want to know how much of an asset is present on a destination chain, read the outstanding representation supply. It is a stock, it nets automatically, and it is unaffected by how many times the same units travelled back and forth.
What inflates volume without moving anyone's position
- Arbitrage round trips. A price gap between chains is closed by moving value across and back, which registers as volume in both directions.
- Operator rebalancing. A liquidity network must continuously move inventory back to the side that is being drained, and that traffic is bridge traffic by any mechanical definition.
What the escrow tells you that volume cannot
A backing ratio below one is the clearest single alarm in this whole area, because it says claims exceed collateral. A representation supply falling to zero on one chain while another chain's representation grows is a bridge migration or deprecation, not new demand.
Step-by-Step Workflow
- Fix a block on each chain and record the heights.
- Read the escrow balance and each representation's outstanding supply.
- Compute the backing ratio and account for any difference in both direction and cause.
- For burn-and-mint, pair events on the message identifier and measure the unmatched tail separately.
- Label the operator's own rebalancing addresses and exclude that traffic from user volume.
- Keep the two directions as separate series and report the net alongside the gross.
- Check whether the asset has more than one representation on the destination chain from more than one bridge.
- Answer presence questions with representation supply, not with cumulative volume.
Worked Hypothetical Scenario
This example is hypothetical. Every figure is invented so the arithmetic can be verified, and none of it describes a real bridge.
Reconciling a lock-and-mint escrow
A lock-and-mint bridge holds one escrow on its source chain and has minted representations of the asset on three destination chains. At a fixed block on each chain, the balances are these.
| Location | Units | Role |
|---|---|---|
| Escrow contract, source chain | 412,500 | Collateral |
| Representation on Chain B | 268,000 | Claim |
| Representation on Chain C | 121,400 | Claim |
| Representation on Chain D | 19,300 | Claim |
| Total representation | 408,700 | Claim |
The escrow exceeds total representation by 412,500 minus 408,700, or 3,800 units, which is 0.92 percent of the escrow. The backing ratio is 412,500 divided by 408,700, or 1.0093. That surplus has three ordinary explanations. Withdrawal fees retained by the contract accumulate there. Withdrawals in flight have already burned their representation while the corresponding escrow release has not yet executed, which pushes the ratio above one for the duration. And tokens sent to the escrow address directly, by mistake or otherwise, add collateral nobody has a claim against.
The same reconciliation when it breaks
Now suppose 60,000 units of representation are minted on Chain C through a compromised message path, with no corresponding lock. Chain C's representation becomes 181,400 and the total becomes 468,700, against an unchanged escrow of 412,500.
| Measure | Normal state | After an unbacked mint of 60,000 |
|---|---|---|
| Escrow | 412,500 | 412,500 |
| Total representation | 408,700 | 468,700 |
| Escrow minus representation | +3,800 | −56,200 |
| Backing ratio | 1.0093 | 0.8801 |
The deficit of 56,200 units is 13.6 percent of the escrow, and backing has fallen to 88.0 percent. Nothing about total bridge volume changed, nothing about the escrow balance changed, and a dashboard tracking only locked value would show a flat line. The invariant caught it because it compares two numbers on two chains that are supposed to agree.
Why ranking bridges by locked value ranks architectures
Take two bridges over the same hypothetical day. Bridge L is lock-and-mint with 40,000,000 units in escrow, carrying 500,000 units of volume. Bridge N is a liquidity network with a 5,000,000 unit pool, carrying 40,000,000 units of volume.
| Measure | Bridge L (lock-and-mint) | Bridge N (liquidity network) |
|---|---|---|
| Reported locked value | 40,000,000 | 5,000,000 |
| Daily volume | 500,000 | 40,000,000 |
| Daily turnover of the balance | 0.0125 times | 8 times |
Bridge L shows eight times the locked value. Bridge N carried eighty times the volume. Neither figure is wrong and neither describes the other bridge's quantity, because an escrow is meant to accumulate and a liquidity pool is meant to turn over.
What Can Make the Interpretation Wrong?
- Summing locked value across designs. Escrowed collateral, structural zero and operator working capital are three different quantities sharing one column heading.
- Treating a liquidity pool as bridged capital. The pool was funded by liquidity providers to enable throughput. Users did not put it there and it does not represent anything they moved.
- Double counting gross volume. Summing both directions counts every round trip twice, and arbitrage generates round trips continuously.
- Missing the return leg. One-directional inflow read as rotation ignores whatever went back the other way in the same period.
- In-flight and failed legs. A burn without a matching mint is either a message in transit or a message that failed, and the two look identical until a timeout.
- Rollup withdrawal queues. A canonical rollup bridge holds escrow on the settlement layer with a mandatory delay before release, so escrow and the rollup balance are out of sync by design for the length of the queue.
- Multiple representations. Several bridges can each mint their own version of the same asset on the same chain. They are separate contracts, not fungible, and summing them overstates what one trade can reach.
- Counting message volume as asset volume. A general message-passing layer carries instructions, and asset transfer is one application among several.
Cross-Network and Provider Comparison
Rollup canonical bridges are a special case. They are lock-and-mint, with the escrow held by the rollup's bridge contract on the settlement layer, and the destination mint expressed as a balance credit inside the rollup. What makes them different is the withdrawal path: an optimistic design imposes a challenge period and a validity-proof design imposes proof latency, so the escrow and the rollup-side balance are deliberately out of sync for the duration of the queue. Reconciling them instantaneously is not possible and a monitor that expects it will alarm constantly.
Non-EVM destinations complicate the burn-and-mint reconstruction, because there may be no event log in the EVM sense to filter. Pairing then depends on instruction parsing or on the bridge's own attestation records, which reintroduces the trust assumption that reading chain data was supposed to remove.
Provider disagreement is dominated by three choices: which designs are included in an aggregate, whether operator rebalancing is stripped out, and whether volume is gross or net. Two providers can agree transaction by transaction and publish figures that differ by a multiple. The DefiLlama data definitions are worth reading specifically for how a large aggregator draws these lines, because the definition is the number.
Advanced Analytical Methods
Continuous escrow reconciliation
Recomputing the backing ratio every block turns a periodic audit into a monitor. Alarm on the ratio falling below one, and treat a small positive surplus as expected rather than as noise to be smoothed away.
Leg pairing with a latency distribution
For burn-and-mint, measure the elapsed time between each burn and its matching mint and keep the distribution. The unmatched tail beyond the usual latency is the population worth inspecting, and a shift in the distribution is an early sign of relay degradation.
Directional decomposition and round-trip identification
Keeping the two directions separate makes net position computable, and matching an outbound leg to a later inbound leg from the same address isolates arbitrage traffic from one-way migration.
Operator address labeling
Rebalancing by the bridge itself is mechanically indistinguishable from user traffic until the operator's addresses are labeled. On a liquidity network this can be a large share of measured volume.
Design-neutral aggregation
The one quantity that aggregates cleanly across all three designs is representation supply per asset per chain. It is a stock and it is design-independent.
Practical Checklist
- I identified which of the three designs this bridge uses.
- I computed the backing ratio and explained any gap in both size and direction.
- I labeled the operator's rebalancing addresses and excluded them from user volume.
- I checked whether more than one bridge issues a representation of this asset on this chain.
- I answered presence questions with representation supply rather than cumulative volume.
Frequently Asked Questions
What does it mean when escrow is smaller than the representation supply?
It means more claims exist than backing, which is the signature of an unauthorized mint or a compromised message path. The direction matters more than the size. Escrow slightly above representation is the normal state, produced by fees retained in the contract, in-flight withdrawals where the destination burn has executed but the source release has not, and tokens sent to the escrow address by mistake. Escrow below representation has no benign explanation.
Why do bridge volume and bridge total value locked disagree so often?
They measure a stock and a flow, and the ratio between them is a fingerprint of the design rather than a health measure. A liquidity network is engineered to turn a small pool over many times a day, so it shows high volume against low locked value. A lock-and-mint escrow accumulates and does not turn over, so it shows large locked value against modest volume. Ranking bridges by locked value therefore ranks architectures.
Can I reconstruct a burn-and-mint transfer from chain data alone?
Only by pairing a burn event on the source chain with a mint event on the destination chain using the message identifier or nonce that the protocol carries between them. Without that identifier the two events are unrelated records on two ledgers. Pairing also leaves an unmatched tail at any moment, because messages in transit have burned but not yet minted, and a failed relay looks identical to a slow one until it times out.
Does high bridge inflow to a chain mean capital is rotating there?
Not on its own, because most published bridge volume is gross and one-directional totals are usually reported without their return legs. Round trips by arbitrageurs and rebalancing by the bridge operator itself both inflate the figure without moving anyone's net position. The measure that answers the rotation question is the outstanding representation supply of the asset on that chain, which is a stock and nets automatically.
What is the difference between a lock-and-mint bridge and a liquidity network bridge?
A lock-and-mint bridge holds the original asset in an escrow contract on the source chain and issues a representation on the destination, so the escrow balance and the representation supply should match. A liquidity network bridge does not lock anything: it pays the user from a pool already sitting on the destination chain and rebalances between pools later. That difference determines what the data shows. Only the first produces a verifiable escrow to compare against supply, while the second produces flows with no corresponding lock to check them against.
Why does a bridge exploit distort a historical flow series?
An exploit typically drains escrow in a small number of very large transactions, and the movements that follow are the attacker dispersing funds and the protocol responding, none of which represent user demand. Left in the series, this appears as an enormous outflow followed by unusual patterns for days. Analyses covering a period containing an incident need it identified and treated separately rather than smoothed away, because the same transactions also break the escrow-to-supply relationship that the rest of the method depends on.
How do you avoid double counting an asset bridged more than once?
By tracking the representation lineage rather than the balances. An asset bridged from chain A to chain B and then to chain C exists as one locked original and two representations, and adding supply across all three chains counts the same economic unit three times. The only reliable approach is to anchor on the escrowed original and treat every downstream representation as a claim on it. Where a representation is itself bridged by a third-party bridge with its own escrow, the chain of claims has to be followed link by link.
Do canonical and third-party bridges produce comparable data?
Not directly. A canonical bridge is the one operated as part of the destination chain's own infrastructure, and its escrow contracts are well known and usually well labelled. Third-party bridges serve the same route with their own contracts, their own representations of the same asset, and often much thinner coverage in public datasets. A flow figure for a chain built only from its canonical bridge understates total movement, and one that mixes bridges without deduplicating their representations of the same asset overstates it.
How does batching affect the timing of observed bridge flows?
Some bridges aggregate many user transfers and settle them in periodic batches, so the on-chain record shows one large movement at the batch time rather than a series of small ones as users initiated them. The total is correct but the timing is not, which matters for any analysis aligning flows to price moves or to a specific event. Comparing a batching bridge against one that settles per transfer, without accounting for this, produces a spurious difference in how flows appear to cluster.
References
These sources should be reviewed during editorial verification. They support data structures and methods, not the hypothetical conclusion. Provider formulas, chain rules, and APIs can change. Confirm current documentation before publication.
- Ethereum.org: Technical Introduction: Contract accounts and stored state, which is what an escrow balance is.
- Ethereum.org: Transactions: Receipts and logs, the records used to pair burn and mint legs across chains.
- Etherscan API Documentation: Address and token lookups used to read escrow balances and representation supply.
- DefiLlama: Data Definitions: How a large aggregator defines total value locked, the quantity most often misapplied here.
- Dune: Data Explorer and Raw Tables: Raw logs and traces needed to reconstruct bridge legs on each chain.