​[HERO PLACEHOLDER: settlement-finality-hero.png]
​

Settlement finality is the point at which a blockchain transaction becomes irreversible. After finality, the payment cannot be reorganized out of history, double-spent, or unwound by a chain reorg. For treasury teams moving stablecoins between exchanges, custodians, and payment partners, finality is the moment risk leaves the books and cash is truly cash.
​

The catch: every chain defines finality differently. Bitcoin asks you to wait an hour. Ethereum delivers it in about 13 minutes. Solana lands in seconds. Arbitrum gives you fast soft finality but a full week for L1 finality. If your AP system treats "transaction confirmed" as "money received," you are exposing the business to timing risk that varies by an order of magnitude across networks.
​

This guide breaks down the four finality models, compares the five chains enterprise treasuries actually use, and lays out the operational implications for payment timing.
​

Answer: Blockchains achieve finality through probabilistic confirmation, economic finality via slashable stake, instant BFT finality, or layered L1-anchored finality. Each model trades latency against the cost an attacker would need to pay to reverse a confirmed transaction.
​

Used by Bitcoin and other proof-of-work chains. A transaction is never strictly final; it becomes exponentially harder to reverse with each block built on top. After six confirmations (roughly 60 minutes on Bitcoin), the probability of a reorg drops below the threshold most exchanges and custodians treat as settled.
​

Probabilistic finality is the oldest model and the most battle-tested. The cost is latency. An hour of waiting is fine for cold-storage transfers; it is unworkable for retail payments or interbank settlement windows.
​

Used by Ethereum after the Merge. Validators stake ETH, and a transaction is considered finalized once two consecutive epochs (about 12.8 minutes) attest to it. Reversing finalized blocks would require an attacker to control one-third of staked ETH, and the protocol would slash that stake. Finality is "economic" because the cost of a reversal is denominated in dollars, not just compute.
​

For treasury teams, economic finality is a cleaner promise than probabilistic confirmation. The window is bounded and the cost of attack is auditable from on-chain data.
​

Used by Solana, Cosmos chains, Aptos, Sui, and most newer L1s. BFT consensus protocols agree on each block before moving to the next. Once a supermajority of validators sign, the block is final. There is no probabilistic tail and no waiting period.
​

Solana confirms in roughly 12.8 seconds for optimistic confirmation and fully finalizes within 30 seconds. Cosmos chains with Tendermint reach instant finality at every block, typically every 1 to 6 seconds. BFT trades a more complex validator set for sub-minute settlement.
​

Used by Arbitrum, Optimism, Base, and other optimistic rollups. L2s have two finality clocks. Soft finality arrives when the sequencer orders the transaction (under a second). True L1 finality arrives only after the seven-day challenge period elapses, during which fraud proofs can claw back invalid state.
​

This split is why bridging out of an optimistic rollup natively takes a week. Fast bridges and intent-based protocols front the liquidity in exchange for a fee, but the underlying L1 finality clock still runs seven days.
​

Answer: Finality determines how long capital is exposed to reversal risk after a payment leaves your wallet. A treasury team that treats a Bitcoin "confirmation" the same as an Ethereum "finalization" is mispricing risk by roughly five times in latency and an unbounded amount in attack cost.
​

Three concrete failure modes show up in production:
​

Payment timing mismatches. If your ERP marks an invoice paid on first confirmation, an attacker who reorgs the chain (or a sequencer that censors a transaction) can leave the customer's wallet credited and yours empty. Waiting for true finality eliminates the exposure.
​

Cross-chain settlement risk. Bridges that mint wrapped assets before source-chain finality have lost hundreds of millions of dollars to reorg attacks. Production bridges now wait for finality on the source chain before releasing on the destination chain. Knowing which clock applies is a vendor-due-diligence question.
​

Working capital lockup. An exchange that requires 60 minutes of Bitcoin confirmations before crediting USDT effectively locks the underlying capital for that hour. On Solana, the same flow clears in under a minute. For high-velocity treasury operations, the choice of settlement rail compounds across thousands of payments per month.
​

Answer: Optimistic rollups like Arbitrum publish transaction batches to Ethereum, then wait seven days before those batches are considered final at the L1 level. During the challenge window, any validator can submit a fraud proof to revert invalid state. Zero-knowledge rollups achieve faster L1 finality (typically hours) because validity is proven cryptographically rather than disputed economically.
​

The seven-day window is not arbitrary. It is calibrated to give honest validators enough time to detect and challenge fraud even under heavy censorship attempts. Most treasury flows that operate within a single rollup never see this window because the rollup's own state machine is final from the sequencer's perspective. The window matters when assets leave the rollup back to Ethereum.
​

Soft finality on Arbitrum is the moment the sequencer orders your transaction. It is fast (sub-second) and reliable in practice because sequencers stake reputation and economic incentives on honest ordering. Hard finality is the seven-day clock. Enterprise contracts should specify which one is required for settlement triggers.
​

For internal accounting, soft finality is usually sufficient. For external settlement (paying a vendor, releasing collateral, closing a hedge), hard finality or a bridged equivalent backed by liquidity provider capital is the safer trigger.
​

Answer: Confirmation is a block-explorer concept: how many blocks have been built on top of your transaction. Finality is a protocol-level guarantee: the network has committed to a state that cannot be reverted without provable cost. Bitcoin uses confirmations because it has no built-in finality. Ethereum has both: 1 confirmation arrives in 12 seconds, but finality requires 2 epochs.
​

The practical implication: do not conflate the two. A Solana transaction with 1 confirmation already has BFT finality. An Ethereum transaction with 1 confirmation does not have economic finality yet. The exchange or custodian's credit policy should reflect this difference, not treat all "1 confirmation" events identically.
​

Finality definitions and Ethereum 2-epoch math: Ethereum.org consensus documentation. Solana optimistic confirmation timing: Solana docs on commitment levels. L2 challenge periods and finality clocks: L2Beat risk framework. Bitcoin probabilistic finality analysis follows the standard six-confirmation convention used by major exchanges.
​