Transaction batching lowers network fees when it replaces several separately submitted transactions with one settlement, because the fixed per-transaction overhead is paid once.

That matters more in 2026: wallets can expose batching instead of making you sign every same-chain step separately. EIP-7702 lets an externally owned account temporarily delegate to contract code, while ERC-5792’s wallet_sendCalls describes a batch request. The same idea powers rollups, where many off-chain actions are posted as one data batch. The saving is real, but it is not a lower gas price.

What the saving actually measures

On Ethereum, the fee is gas used multiplied by the effective gas price. Every standalone transaction carries intrinsic overhead; a plain ETH transfer alone uses 21,000 gas, before contract work or calldata. A batch pays one outer transaction and one signature-and-nonce envelope, then pays for each inner call and its calldata. The saving is the repeated envelopes, not the work itself.

That explains why a three-step flow—approve, deposit, then claim—can cost less as one wallet batch, but not by three times. Multicall encoding, contract execution, and failure handling add gas. If calldata or a busy fee market dominates, the percentage saved shrinks.

Where this changes a bridge decision

There are two different batches. A smart-wallet batch combines your same-chain actions. A rollup batch combines many users’ actions before data is published. On Manta Pacific, the second batch spreads data-availability and settlement costs across users; it does not erase the source-chain transaction, a destination-chain action, or a bridge provider’s fee.

Compare the amount received, not only the wallet’s gas line. A route through Symbiosis Finance or Owlto Finance may reduce friction, yet its relayer, liquidity, and protocol charges still belong in the total. If the alternative is a direct route into Manta Pacific, Manta Bridge is the route to compare.

When to batch

Batch when two or more calls happen on one chain, the wallet and contracts support atomic execution, and the delay is acceptable. Send separately when calls must settle independently or when the batch’s calldata and execution overhead approach the repeated transaction cost. The verdict is simple: batching saves fixed network overhead; it does not make cross-chain execution free.