Ethereum researchers have proposed cutting transaction finality from about 16 minutes to roughly four minutes, according to a draft filed Wednesday. The unfinished plan has no activation date, but its initial target is a fourfold improvement in the time it takes for a transaction to become final.
The draft was filed by Ben Edgington, who coordinates Ethereum's fast-finality research, along with fellow researchers Francesco D'Amato, Roberto Saltini, Yann Vonlanthen and Mikhail Kalinin.
Background on finality
In Ethereum's current mechanism, a transaction can land in a block well before the network treats it as final—the point where validators, the computers that secure the network, commit to never rewriting it. Anyone moving funds off Ethereum has to decide when it is safe to release money on the other side. But sixteen minutes is a long time to have money in limbo.
For example, if someone sends $10,000 from Ethereum to another network, a transfer service can pay them right away from money it already holds there, then wait to get repaid once the original transaction settles. The customer gets funds in seconds, but the service has $10,000 tied up until Ethereum is done.
Ethereum Foundation research counts that idle money as a real cost for these firms, and a shorter wait lets them reuse it sooner. In practice, a service fronting a steady $1 million an hour would have about $267,000 waiting on repayment with a 16-minute delay. At four minutes, that drops to roughly $67,000.
Existing fast options
Stablecoin providers such as Circle, which issues the dollar-pegged USDC, already charges for speed. Its fast route from Ethereum currently costs 0.01%, or $1 per $10,000, before network and application charges. Its standard route carries no Circle transfer fee. Circle also caps the total value of fast transfers waiting on final settlement, and each completed settlement frees up room for new ones.
Splitting the vote
The 16-minute wait comes down to how Ethereum's validators vote. Each one casts a single vote that does two jobs at once: helping pick the next block and helping lock older blocks in for good. Validators take turns casting those votes, and a block only becomes final after two full rounds of everyone voting. Both jobs share one vote so finality can only move as fast as that rotation.
The proposal gives each job its own vote, so the locking-in can run on a faster clock that developers can keep tightening over time. The researchers say splitting the votes would also make it easier to eventually swap in quantum-resistant signatures. Ethereum would need that upgrade if quantum computers ever get strong enough to break today's cryptography.
However, four minutes would still be slow next to Solana, which switched on its Alpenglow upgrade Wednesday with a target of finality in about 150 milliseconds.
Unresolved details
The draft leaves voting timings, penalties and parts of its security analysis unresolved as of Thursday. Edgington wrote that "many aspects remain TBD" and that the immediate aim was getting a proposal number—which puts the idea into Ethereum's formal upgrade pipeline.
