The Ledger
Zcash Proposal Targets 20-Second Blocks for NU7.1
Dev Ojha proposes cutting Zcash’s block target spacing from 25 seconds to 20 seconds and wants the change included in NU7.1. The proposal reports stress-net orphan rates of 1.9% and 3.3% under two network configurations.

On 2026-11-08, Dev Ojha proposed reducing Zcash’s block target spacing from 25 seconds to 20 seconds and said he wanted the change slated for inclusion in NU7.1. The proposal describes the change as latency-only and asks for it to be included in the next poll. It presents the proposal as a follow-on to ZIP 218, which lowered block time to 25 seconds in NU7.
Proposed latency changes would keep the 2 MB limit
The proposal says shorter spacing would reduce the wait for a first confirmation and for back-to-back payments, exchange deposits and cross-chain swaps such as NEAR Intents. It would keep the block size limit at 2 MB and scale ZIP 218’s shielded action limits by four-fifths: the global shielded budget, Orchard limit and Ironwood limit would each move from 330 to 264, while the Sapling limit would fall from 300 to 240 inputs and outputs per block. The draft says scaling those limits with block time keeps daily maxima exactly the same, apart from up to a 76kb/day difference due to more block headers; leaving them unchanged would raise the worst-case load by 25%.
Stress-net results and the earlier safety bar
In the authors’ tests on an 80-miner network spread across 12 regions, full 2 MB blocks at 20-second spacing at the worst case—maximum Ironwood load, then transparent afterwards—produced a 1.9% orphan rate on a Zakura-only network and 3.3% on a network split evenly between Zakura and Zebra-Andele. The proposal compares those results with the 4.68% empirical orphan rate and 3.26% theoretical rate cited in the earlier work used to assess 25-second blocks. Its authors say the measurements meet that earlier readiness bar and say they believe mainnet results would be far better because miners and pools are geographically clustered.
Confirmation policy and node requirements
The draft says it does not ask services to reduce confirmation counts. It argues that services setting confirmation policy by wall-clock time or accumulated work need no change, and says node CPU requirements would not increase; bandwidth under sandblasting would rise by 25%, with 100kb/s needed to stay synced. The draft addresses the question raised by @shielded-nate, “Doesn’t a shorter block time reduce security per confirmation?” and says the effect depends on the threat model.
Zeck Zodel questions the 20-second target
On 2026-11-09, Zeck Zodel questioned whether 20 seconds is the best target, asking how the choice compares with 17 seconds or 21.4 seconds. Ojha said the proposal reflects the available analysis and described the goal as iterating on block time and block size while remaining more than safe against known attack vectors. He said the chain processes 80KB/s at peak today and that the proposed change would bring that to 100KB/s; he suggested the next big win after 20-second blocks should be doubling maximum block size, which he said can double shielded throughput with Tachyon.
Sources
- Original source Zcash Community Forum forum.zcashcommunity.com