Why Finality on L2 Chains Takes Longer Than You Think
On an L2, transactions feel instant — you sign, your wallet shows success, and balances update right away. But when you bridge those funds to another chain, things don’t always move as quickly. Sometimes the transfer stalls for minutes, even though everything looks “done” on the source chain.
This is the same principle centralized exchanges use. When you deposit crypto, they see your transaction immediately — but you still wait for a certain number of confirmations before the funds are credited. The exchange needs to be sure the deposit cannot be reversed. Bridges follow the same rule.
The reason is finality. Before Allbridge Core can release your tokens on the destination chain, it must be absolutely certain that your transaction on the source cannot be reversed. This waiting period is what protects cross-chain transfers from risks like reorgs or batch delays.
From the user’s perspective, that extra time may feel like lag. In reality, it’s the safeguard that makes sure once your funds arrive, they stay there.
What Finality Really Means
Finality is the point where a blockchain transaction becomes irreversible. Once it’s final, no competing chain or validator set can undo it.
There are two main ways blockchains reach this state:
- Probabilistic finality – common in proof-of-work systems. A transaction grows safer the deeper it sits under new blocks, since reorgs become less likely.
- Deterministic finality – used in many proof-of-stake and modern consensus systems. Once validators confirm a block, it’s locked forever with no chance of rollback.
The reason finality matters is that blockchains can briefly produce forks — competing versions of history. This happens if two miners or validators propose different blocks at the same height. The network eventually chooses one chain and abandons the other, which erases any transactions that were in the discarded block. Finality is the moment after which such reorganizations are no longer possible.
Note: You can look for reorgs or forks history of a particular blockchain on block explorers. Here, for example is reorgs page on Etherscan for Ethereum.
Finality on L1 vs. L2
On Layer 1 blockchains, finality depends on the consensus design:
- Ethereum finalizes after two epochs (around 12–15 minutes).
- Solana reaches finality within seconds thanks to frequent block production.
- Sui achieves near-instant deterministic finality through its consensus protocol.
- Stellar locks transactions in every few seconds via its federated voting system.
Layer 2 chains work differently. Transactions appear instant, but they’re not final until the L2 posts a batch of transactions to Ethereum and that Ethereum block itself finalizes. This two-step dependency means L2 finality almost always takes longer than on most L1s.
Note: You can monitor batches written to L1 from most of L2 block explorers. For example, here is such a page on Arbiscan for Arbitrum.
And in practice, that delay can stretch further. Batches may wait in the Ethereum mempool if gas prices spike, or get slowed by sequencer backlogs when traffic is high. Some L2s have even paused batch posting entirely during peak load or maintenance.
These issues aren’t just theoretical. Arbitrum has seen hours-long gaps between batches, Optimism has slowed during spikes, and in 2023 the craze for EVM “ordinals” inscriptions clogged multiple L2s, delaying state updates and finality. In every case, users saw “confirmed” transactions on the L2, but true finality — and cross-chain settlement — only arrived once Ethereum finalized the corresponding batches.
What It Means for Allbridge Core Users
For bridges, there is no safety net: once funds are paid out on the destination chain, they cannot be reclaimed. If the source transaction were later rolled back, the bridge would absorb the loss.
That’s why Allbridge Core only releases funds after the source-side transaction is truly final — including both the L2 batch posting and Ethereum’s own finalization. This sometimes makes L2-origin transfers slower than L1 transfers, and delays can stretch further during congestion or sequencer issues.
The tradeoff is simple: a bit more waiting time, in exchange for certainty that every transfer is secure and irreversible.
