Zcash has moved Network Upgrade 7 onto public testnet, putting the privacy chain’s next consensus change under live conditions ahead of a November 2026 mainnet target.
The Zcash Foundation shipped Zebra 7.0.0-rc.0, the first release candidate for NU7, on October 2, 2026.
That build is written to activate the upgrade on the public test network at block height 4,465,026, a point the teams expected around October 6.
Operators of testnet nodes were told to update before that height if they want to remain on the NU7 chain and help exercise the rules under real traffic.
Mainnet operators were told not to act yet: no production activation height has been locked in, and the package is still a candidate rather than a stable release.
The schedule was set in mid-September after sentiment collection and talks among the Zcash Foundation, Project Tachyon, Valar, ZODL, and Shielded Labs.
Developer Sean Bowe reported unanimous agreement on scope and timing.
The checklist called for feature completion by September 30, testnet activation on October 6 at the height agreed by the Zakura and Zebra teams, a final mainnet decision on October 20, and a provisional production activation on November 5.
That November date stays contingent on how the test network behaves.
The October 20 review, scheduled about a week before Zcon, is where engineers are due to choose a mainnet height or delay the cutover if testing shows problems.NU7 is narrower than some earlier proposals.
It does not introduce a new transaction format, and the teams have said ordinary wallets should see little change.
Full nodes, indexers, and block explorers are more likely to need updates.
The headline protocol change, specified in ZIP 218, shortens the target interval between blocks from 75 seconds to 25 seconds.
That interval has been in place since Zcash launched in 2016.
A first confirmation would therefore arrive, on average, in about a third of the time it does today, which matters for point-of-sale payments, exchange deposits, and bridges that wait for a fixed number of confirmations.
Daily issuance is meant to stay the same: remaining halving intervals are stretched by a factor of three, each scheduled block subsidy is divided by three, and funding-stream periods are adjusted to match.
The 21 million ZEC cap and the familiar four-year halving pattern are both preserved.A second piece is the Network Sustainability Mechanism.
Under the design in ZIP 235 and ZIP 237, 60 percent of each block’s transaction fees is taken out of circulation and held in a reserve, while miners keep 40 percent so they still have a reason to include user transactions.
Later blocks reissue a small, predictable slice of that reserve on top of the normal subsidy.
Public reserves are seeded with the historical shortfall from before NU6.
Coinholder polling backed keeping halvings and pointed to February 2031 as the start of fee reintroduction.
On testnet, the adjusted third halving sits at height 4,497,948, and reserve reissuance begins at 7,305,222.
Faster blocks also raise shielded capacity while trying to limit spam.
NU7 sets per-block caps of 330 Orchard or Ironwood actions, 300 Sapling inputs and outputs, a shared budget of 330 across pools, and no new Sprout JoinSplits.
The Foundation says those limits more than double Orchard protocol throughput and cut the worst-case sync load a spammer can push onto light wallets by about 37 percent.
Zebra applies the same ceilings when building templates and when admitting transactions to the mempool.
Version 4 transactions are disabled as part of the upgrade.
Because the coinbase now depends on selected fees and on the parent block’s reserve, mining templates list only time as mutable; pools are expected to request a fresh template rather than rewrite the transaction set themselves.
Zebra’s state database moves to format v29, which stores the new reserve, and a compatible v28 database is migrated automatically on first start.
Downgrades are not supported.
Testnet configs that pin old consensus values can prevent the node from starting until those overrides are removed.
The testnet window is deliberate.
Consensus, mining, wallets, and node storage all change, so the candidate release is meant to surface issues before production. A final mainnet height is still to be set after that review, with November 5 remaining the working target rather than a guaranteed activation.