Ethereum researchers are developing a new consensus architecture that could reduce the network’s economic finality time from roughly 13–19 minutes to seconds.
The proposal separates the work of producing blocks and selecting the chain head from the process that gives transactions economic finality. A smaller, rotating validator committee would handle a faster “heartbeat” chain, while a separate finality mechanism would continue to draw on Ethereum’s wider validator set.
The work remains under development. The Ethereum Foundation says specifications and prototypes are underway and describes decoupled consensus as the leading headliner candidate for the proposed I* fork. No finality upgrade has been formally assigned to a fork, however, and seconds-level economic finality has not been deployed on Ethereum mainnet.
Why Ethereum’s Current Finality Takes So Long
Ethereum creates a new slot every 12 seconds, but its current Gasper consensus system does not make blocks economically final that quickly. Gasper combines LMD-GHOST for chain-head selection with Casper FFG for economic finality.
Casper FFG requires two voting rounds around checkpoints, and an epoch contains 32 slots. That produces a minimum finality time of about 768 seconds, while the expected wait for a randomly submitted transaction is about 960 seconds. Under normal conditions, the practical finality window is roughly 13 to 19 minutes.
The delay leaves a portion of the recent chain without full economic finality. The Ethereum Foundation estimates that about 63 to 95 recent blocks can remain unfinalized under normal conditions.
Researchers are trying to remove the need for block production and finality to operate within the same latency constraints. The Ethereum Foundation’s research on upgrading finality describes decoupling the two functions as the central architectural change.
A Smaller Committee Could Handle the Fast Path
Under the proposed design, a relatively small committee would handle block production and fork choice, with research examining committees of around 256 or 512 validators. The finality mechanism would operate separately rather than requiring the same voting process to handle every part of consensus.
The approach is intended to reduce the communication and processing burden of the fast path without replacing Ethereum’s broader validator population with a permanently small committee.
Researchers have examined Goldfish, a fork-choice design that uses vote expiry and view-merge to improve resistance to reorganizations while operating with smaller committees. The broader design also needs to remain available as validator participation changes, including when some validators go offline and later return.
This work builds on several earlier attempts to accelerate Ethereum finality, including Single-Slot Finality research published in 2023 and 3-Slot Finality research in 2024. The 2025 Lean Ethereum proposal also described a consensus system capable of finality in seconds.
The 12-Second Fast Confirmation Rule Is Different
Ethereum researchers have separately studied a Fast Confirmation Rule that can identify blocks as effectively stable after a single 12-second slot under specified network assumptions.
A study by Roberto Saltini, Mikhail Kalinin, Luca Zanolini, Francesco D’Amato, Aditya Asgaonkar and Chenyi Zhang evaluated the rule using about 26 hours of Ethereum mainnet data beginning with block 12,607,808 on September 17, 2025.
The study reported that 94.67% of the observed blocks could be confirmed within one slot, 98.87% within two slots and 100% within eight slots. The researchers also noted that the observation period did not include an adversary deliberately targeting the confirmation mechanism.
Fast confirmation should not be treated as equivalent to Ethereum’s full economic finality. The current Gasper finality process remains in place, while the broader seconds-level effort seeks to redesign that underlying process.
Faster finality has drawn particular attention from bridges, interoperability systems and other infrastructure that must account for the period before Ethereum transactions receive stronger settlement guarantees. In its 2026 stakeholder research covering 19 organizations and individuals, the Ethereum Foundation found that 10 to 20 seconds was frequently viewed as a useful target for bridges, interoperability and intent-based systems.
The same research found that finality below one minute could provide meaningful benefits for many applications. Systems that need responses in roughly one to four seconds may still need preconfirmations or Layer 2 mechanisms rather than relying on Ethereum mainnet finality.
Researchers have also considered reducing finality to a single voting round, but the proposal involves a security trade-off. Depending on the design, the amount of adversarial stake Ethereum could tolerate could fall from roughly one-third to around one-fifth or one-sixth. Feedback collected by the Ethereum Foundation showed limited support for weakening the finality safety margin simply to reduce latency.
The current research therefore focuses on changing the architecture rather than relying on a single shortcut. Validator consolidation, attestation aggregation, networking improvements and new consensus mechanisms are being examined alongside the separation of the fast heartbeat from the finality layer.
Discover more from Aree Blog
Subscribe now to keep reading and get access to the full archive.
