Everyone is quoting 25 seconds. The number that decides whether this upgrade matters is 60. And the number after that isn't a number at all — it's a date, and it's the only line in the release that can cost someone their entire position.
Zcash's seventh network upgrade, NU7, compresses the block interval from 75 seconds to 25 seconds. Three confirmations drop from roughly three minutes and forty-five seconds to about seventy-five. It's a clean, quotable improvement, and it has been running on testnet for months while the node software catches up.
But NU7 ships two other changes in the same activation. Sixty percent of the fees collected in every block get pulled out of circulation, with a scheduled return to miners beginning around February 2031. And Sprout — the original shielded pool, live since 2016 — gets switched off. Not deprecated. Off. Any ZEC still sitting inside it at activation stops being spendable.
I have spent nine years reading upgrade notices for a living. The performance line is the one that earns a press cycle. The deadline line is the one that takes people's money.
What a Zcash network upgrade actually commits to
Zcash doesn't ship changes the way application chains do. A network upgrade is a bundle of ZIPs, agreed in advance, activated at a specific block height. Every node that wants to stay on the network must be running software that knows what happens at that height. If you don't upgrade, you don't fork — you just stop being on the chain that has economic value. That's the design. It's brutal and it's clean.
NU7 is the seventh of these. The interval change is the headline ZIP. The fee treatment and the Sprout retirement ride along in the same activation, which means the chain gets three economically distinct changes at one block height, and the market will only talk about one of them.
To read any of this you need the pool structure in your head. Zcash has three shielded pools — Sprout, Sapling, Orchard — plus transparent addresses. Value moves between them through a turnstile: you can go from transparent to shielded, or from an older pool to a newer one, but the accounting is public even when the contents aren't. Every pool's balance is auditable by anyone running a node. That single property is why the Sprout shutdown is a verifiable event rather than a press statement, and it's why I care about this upgrade at all.
The subsidy schedule matters too. Zcash halves its block reward every 1,046,400 blocks. At 75-second spacing that's roughly four years, which is why the last halving landed in late 2024. Hold that number. It's about to become the most interesting arithmetic in the upgrade.
And the node layer is mid-migration. The original reference implementation is being retired in favour of Zebra, which means the coordination cost of a consensus change right now is higher than it has been in years. "Running on testnet, awaiting final integration" is not a scheduling detail. It's the part of the process where things slip.
The interval: three times faster, and four things that change with it
Seventy-five to twenty-five is a factor of three. Run it through the confirmation math and the user-facing story is real: a three-confirmation wait falls from 225 seconds to 75. For a merchant terminal or a settlement desk, that's the difference between "we'll hold your order" and "done."
But the interval is a parameter that touches everything downstream of it, and the downstream effects rarely make the announcement.
First, the emission schedule has to be re-indexed, or it breaks. The halving interval is denominated in blocks, not seconds. At 25-second spacing, 1,046,400 blocks elapse in roughly 302 days, not four years. If NU7 doesn't reparameterize the interval, the subsidy schedule compresses by a factor of three and the next three halvings arrive inside two and a half years. That is not a cosmetic bug. That is a monetary policy event disguised as a performance upgrade, and it needs an explicit ZIP line saying so. Watch for it. If it isn't there, ask why.
Second, coinbase maturity gets shorter in wall-clock terms. Zcash requires 100 confirmations before mined coins are spendable. At 75 seconds that's a little over two hours. At 25 seconds it's roughly forty-two minutes. Mining pools that batch payouts on a maturity trigger will pay out three times as often, or will need to change their batching logic. Exchanges that credit deposits on internal confirmation thresholds will silently become three times more permissive. Nobody will announce that either.
Third, propagation latency stops being negligible. A block takes one to two seconds to reach most of the network. Against a 75-second interval that's under three percent of block time. Against 25 seconds it's four to eight percent. Stale rates scale roughly with that ratio, and stale blocks are unpaid work. For a chain whose hashrate is already thin and concentrated, tripling the stale rate is a quiet transfer of income away from small miners toward whoever has the best network position. That's a centralization vector produced by a speed upgrade, and it deserves a paragraph in the ZIP that it probably won't get.
Fourth, blockspace triples and almost nobody needs it. Zcash blocks are capped at 2MB and are mostly empty. Tripling the number of blocks triples theoretical throughput, which matters only if the blocks are full. They aren't. Volume is noise; token velocity is the heartbeat. If the chain isn't processing demand, making it three times faster at processing nothing doesn't change the economics — it changes the dashboard.
There's a fee-market consequence buried in there. More blocks means more blockspace per unit time, which means lower fee pressure per block for any given level of demand. A chain that just tripled its supply of blockspace is a chain that made its own fee market cheaper. Which brings us to the second change, and the reason it exists.
The 60 percent: a burn that is mostly a promise
Sixty percent of the fees collected in each block are removed from circulation. The remainder is distributed as normal. The removed portion has a return path to miners, scheduled to begin around February 2031.
Start with the size of the number being burned today. The standard Zcash transaction fee is a fraction of a coin — a rounding error against the block subsidy. A block with a few dozen transparent transactions collects a fee total you could lose in a wallet sweep. Sixty percent of that is dust. On any given day in this market, the burn is economically invisible.
That's precisely why it's worth reading carefully. A mechanism this small isn't designed for today. It's designed for a future where fees matter, and the design choice tells you what the protocol expects.
Look at the date. February 2031 sits just ahead of the next subsidy halving, at the point where the block reward is at its thinnest and the security budget starts to lean on transaction fees instead of issuance. A fee stream that accrues off-balance for five years and then unlocks to miners is a pre-commitment device: it builds a fee-funded security budget during the years when nobody would voluntarily fund one, and it releases it exactly when the subsidy stops carrying the chain. That is a coherent piece of monetary engineering, and it is a far more interesting fact about NU7 than the block interval.
The engineering is only coherent if one question gets answered precisely: removed from circulation how? Escrowed and burned-and-reminted are not the same thing. If the 60 percent sits in a provable escrow address, total supply is unchanged and only float moves. If it's burned and later re-issued, supply moves twice and the emission curve anyone modelling Zcash needs to be rebuilt. Both are defensible. They produce different numbers. Anyone building a supply model off this upgrade should demand the ZIP language rather than the summary.
And be honest about the dependency. A fee burn on a chain with negligible fees is a mechanism waiting for a use case. It is a bet that shielded transaction demand returns. If it doesn't, the 2031 unlock is a rounding error with a calendar entry.
Sprout: the shutdown is the only irreversible line in the upgrade
Sprout is the original pool. It went live with the chain in 2016, before Sapling, before Orchard, before anyone had shipped a shielded transaction at scale. It has known limitations and it has been on a retirement path for years. NU7 finishes the job: at activation, the pool closes.
Read the consequence precisely, because the wording matters. Funds still in Sprout at activation do not get burned, seized, or redistributed. They become inaccessible. There is no recovery path, no governance vote, no exception. A private key that controls a note in a pool the network no longer validates is a string of characters with no meaning.
Who is still in there? The pool balance is public — you can query it from any node — and the answer is a small, frozen number that has barely moved in years. That flatness is the tell. A pool whose balance doesn't move isn't a pool with users. It's a pool with keys nobody holds anymore: laptops in drawers, seeds on paper in a storage unit, exchange custodial leftovers from the 2017 cycle, wallets generated by software that no longer exists and restored by nobody.
I've run this pattern before. In 2017 I traced a token migration contract out of Estonia that was draining retail holders — fourteen exchanges, a mapped wallet graph, about $2.5 million moved before I published the contract vulnerabilities and the wallet paths. The mechanism wasn't clever. It was a deadline plus a population that doesn't read. Every rug pull has a trail of paid gas — and the inverse is what makes Sprout dangerous: the funds that never move leave no trail at all. The absence of transactions is the data point, and it's the one nobody watches.
There's a second-order risk that deadline upgrades always create and that nobody budgets for. A migration deadline is a phishing surface. The moment "move your funds before activation" becomes public knowledge, someone starts sending emails about it. Old wallet holders are the perfect target: they've been inactive for years, they're technically rusty, and they've just been told their money is at risk by a clock. Expect fake migration tools, expect fake support channels, expect a wave of them in the four weeks before activation.
If you might be exposed, the check is mechanical. Any wallet created before Sapling existed, restored from a seed generated in that era, or untouched since, should be opened and inspected on current software before activation — not after. Old libraries may not even display a Sprout balance, which is its own trap: a zero on screen is not proof of a zero on chain. And the confirmation signal for everyone else is simple and public: watch the Sprout pool balance drain to zero. When it hits zero before activation, the upgrade is clean. If it's still non-zero at the activation height, that number becomes the permanent record of what was lost.
The contrarian read: latency was never the binding constraint
Here is what bothers me about the entire conversation. Zcash is not losing because three confirmations take four minutes. It is losing because access to it is being narrowed, and none of NU7's three changes touches that.
A privacy chain's constraint is not throughput. It's venue support. One top-tier exchange removed its Zcash spot pairs in 2024; another appended a monitoring tag that functions as a slow-motion delisting warning. Liquidity concentrates in fewer books, spreads widen, and the asset becomes structurally harder to hold for anyone with a compliance function. You can cut confirmation time to five seconds and it will not put a single pair back on a venue that has decided the asset is a liability.
And the legal weather around privacy tooling has not improved. The precedent set when a sanctions authority treated a published smart contract as a sanctioned entity still stands, and it still means that writing open-source privacy code is an activity with legal exposure attached. That precedent is the single largest risk factor for any chain whose core product is a shielded pool, and it is not addressable by a consensus parameter. We followed the ETH, not the promises — I wrote that during the 2021 wash-trading work, when I mapped fifty thousand OpenSea transactions and found clusters of wallets funded from a single source inflating eight million dollars of fake volume. The lesson generalizes: the story is always in the flows and the access, never in the feature list.
So treat the interval change for what it is — a genuine engineering improvement that removes a real friction, in a market where friction is not what's killing the asset. And resist the reflex to model NU7 activation as a catalyst. Upgrades don't move flows. Flows move flows. Capacity upgrades get absorbed and re-priced; I watched post-Dencun blob capacity get eaten by demand within a year and the fee market settle right back where it started. Zcash's blockspace will behave the same way the moment anyone actually wants it.
In a bear market this distinction stops being academic. Survival isn't determined by who has the fastest confirmation. It's determined by who still has venues, who still has miners, and who still has users who can reach their own funds.
What to watch
Three numbers, in order of how much they can hurt you.
First, the emission ZIP. If NU7 doesn't explicitly re-index the 1,046,400-block halving interval, the subsidy schedule compresses by three and the security budget narrative changes entirely. That's the line to read before the line anyone is quoting.
Second, the Sprout turnstile balance. It's public. Watch it go to zero before the activation height. If it doesn't, the remainder is a permanent monument to people who didn't get the memo.
Third, the first thirty days of fee accounting after activation. Sixty percent of a small number is still a small number, but the accounting either reconciles or it doesn't — and how the protocol books that escrow tells you whether the 2031 unlock is a security budget or a press release.
The block interval is the part of NU7 that will be remembered. The deadline is the part that will be felt. Which one do you think the people still holding Sprout notes are reading about tonight?