Block 3,543,405. That's where Zcash flipped a switch early Friday, November 6, and became a different machine from the one that existed the day before. The chain now aims for a block every 25 seconds instead of every 75, and it has started pulling 60% of transaction fees out of circulation to pay for security years from now.

The Zcash NU7 upgrade is a bundle, and the bundle is more interesting than any single item in it. Faster blocks, a fee-recycling scheme, a retired privacy pool and a cap on how much shielded data a block can carry all landed at once. Each trades something for something else.

What actually switched on

The project's official account announced that NU7 had activated on mainnet, adding that "A new era of Zcash is here." Marketing aside, the technical changes are specific.

First, the Network Sustainability Mechanism (NSM). Under it, 60% of each transaction fee is withdrawn from circulation and later returned to miners through block rewards. Payouts don't begin until 2031, and they layer onto the existing halving schedule along a smooth curve. The 21 million ZEC cap stays intact.

Second, the block target. Decrypt reported, citing the Zcash Foundation, that the target interval falls from 75 seconds to 25.

Third, Sprout is gone. That's the original 2016 shielded pool, and about $16 million worth of coins remain locked inside it, now out of reach.

Fourth, a ceiling on how much private transaction activity each block may hold. The aim is to reduce the volume of material a wallet has to sift through before it spots its own incoming payments.

The fee problem Bitcoin left behind

The NSM exists because of a question Zcash inherited from Bitcoin: when the subsidy per block keeps halving, will fees alone pay miners enough to keep the chain secure? Nobody knows. Bitcoin has until roughly the next century to find out in earnest. Zcash, with a similar issuance shape and a smaller fee market, has less patience for hypotheticals.

The design is a clever bit of accounting. Rather than print new coins to top up rewards, which would break the supply cap, the protocol takes a slice of what users already pay and holds it back, then pays it out to miners later. Fees today become subsidy tomorrow. Nothing new is created, only deferred, so the cap holds.

The lag is long. Reissuance begins in 2031, five years from now, so until then the NSM is purely a drain on circulating supply and a promise. Whether 60% is the right fraction was a judgment call by the developers, and I haven't seen a rigorous argument that it's optimal rather than merely defensible. Five years is plenty of time to find out.

Shielded Labs gets half of what it wanted

Jason McGee of Shielded Labs, the organization behind the mechanism, called the activation a huge milestone. He also said the NSM was the group's first project, and that it grew out of an idea developer Nate Wilcox floated years ago under the name Zcash Posterity Fund.

What shipped isn't what Shielded Labs first drew up. Its original design would have scrapped halvings in favor of a smooth issuance curve. Community panels backed that; coinholders didn't. A second design was built that keeps the halvings, and that one is now live.

McGee was candid about it: "We still think the original design was simpler and better." He added that he respected the process that produced the compromise, and that changing the protocol should be hard and resistant to capture by any single group or individual.

That's a gracious thing to say after losing a vote, and a fair description of a governance system working as designed, if slowly. With panels on one side and coinholders on the other, either body can veto a proposal the other loves, and this time coinholders did. The result is a patched design, not a stalemate.

Why 25 seconds is the risky part

This is where the numbers matter most, so here is the arithmetic.

At 75-second targets, the chain produces about 1,152 blocks a day. At 25 seconds, it's 3,456. That's three times as many chances for two miners to find a valid block at nearly the same moment. When it happens, the network keeps one and discards the other. The loser is an orphan block, and every orphan is wasted work.

Shorter intervals leave less time for a block to spread before the next one arrives, so rivals collide more often. No protocol upgrade repeals that.

The team behind Zakura, which last month put forward a fix to protect shielded swaps against quantum attacks, ran a stress test to put a number on it. They used 80 nodes spread across 11 regions and filled every block with private transactions, the heaviest and slowest load Zcash handles. The orphan rate came out at 3.43% with 25-second blocks, against 1.57% at 75 seconds. That's about 2.2 times higher, a smaller jump than the threefold increase in block count would suggest.

Those are worst-case conditions on purpose. Zakura's developers expect the live network to land under 1%, for two reasons: real mining is far more concentrated than a test spread across a dozen regions, and real blocks are far emptier than a stress test's. Concentration cuts both ways, though. A handful of large pools can sync with each other fast, which keeps orphans down, but a privacy coin is hardly going to put that on a brochure.

The payoff for accepting the risk is speed. At 25 seconds, six confirmations take two and a half minutes instead of seven and a half. For a currency whose users put real effort into shielded transactions, shaving that wait is a genuine usability gain. Still, the 1% figure is a forecast, not a measurement, and the first few weeks of live data will say more than any simulation.

The per-block cap on shielded activity looks like a companion move. Faster blocks carrying unbounded private payloads would strain both propagation and wallets. Limiting the data per block holds down scanning costs and keeps each block lighter to relay, which in turn helps the orphan rate. If the cap pinches during busy periods, users will find out the hard way.

Sprout's quiet burial

Retiring Sprout is the least glamorous change and probably the most defensible. The pool dates to Zcash's 2016 launch and predates the cryptographic improvements that later shielded pools use. Maintaining old code for a legacy pool is a standing security liability.

The cost is that roughly $16 million in coins stays locked inside it. Those holders had years to migrate, and presumably some of the coins sit behind lost keys or abandoned wallets. Even so, a sum that size is a reminder that protocol cleanups have owners somewhere, even if they've gone silent.

The Dev Fund argument is next

NU7 doesn't end the political calendar. Zcash is now in the middle of a dispute over funding for Zcash Community Grants, which draws on the Dev Fund, the portion of each block reward earmarked to pay for development work.

Organizing the coinholder vote is a joint job for five groups: ZODL, the Zcash Foundation, Valar Group, Shielded Labs and Project Tachyon. They chose a compressed timetable, saying tensions were high and a drawn-out debate could deepen divisions. Proposals are due November 8. Coinholder and community-panel polling takes place between November 16 and November 30. Should the groups find broad agreement, they plan to back a follow-up upgrade, NU 7.1, to put it into effect.

There's a tie to the NSM here. Among the ideas on the table is an obligation for the grants program to commit 50,000 ZEC of its treasury to the mechanism, so the money that funds ecosystem grants could end up seeding long-term miner pay.

Two things to watch, then. The first is the orphan rate on the live chain over the next month. If it holds near or below 1%, Zakura's forecast looks good; if it creeps toward the 3% stress figure, expect miners to complain loudly. The second is the vote. Coinholders already overruled the panels once on issuance design, and nobody has said whether they'll do it again on Dev Fund money. The poll closes November 30, and the result shapes what NU 7.1 contains.