Author: Blockworks
Compiled by: AididiaoJP, Foresight News
Summary
In the third quarter of 2026, Zcash implemented the first step of its scaling roadmap. The new full node Zakura, released in mid-July, compressed full-network sync time from about 21 hours on Zebra to about 4 hours; nodes started from a pruned snapshot ran in less than two minutes. At the end of July, Ironwood (NU6.3) activated, bringing an alternative shielded pool whose supply reliability is backed by machine-verified mathematical proofs. Holders migrated quickly, with the shielded supply share rising from about 26% to about 29%, and the total amount of shielded ZEC growing about 12% to about 4.9 million coins.
Zakura is jointly advanced by Valar Group and Project Tachyon. Valar Group is led by Osmosis co-founder Dev Ojha, and Project Tachyon is led by Zcash co-founder Sean Bowe. Its design target is about 50,000 private transactions per second, achieved in phases: the node itself, Tachyon's recursive proofs, and Valar's wallet-facing private information retrieval. The Zakura Common cryptography suite released at the end of August increased mobile proof generation speed by more than 14x, reducing proof generation for a standard shielded transaction on an iPhone from more than 3 seconds to about 200 milliseconds.
Holders then participated in governance through the new pool. Nearly 2.4 million ZEC, about two-thirds of eligible supply, participated in an encrypted vote on NU7. With 99.9% support, they chose 25-second blocks, retained the halving schedule, and determined the terms of the Network Sustainability Mechanism (NSM). Mainnet activation is planned for early November. A separate vote approved about $8.4 million in grants from the protocol holder-controlled fund, including about $1.8 million to Valar Group and about $1 million to Project Tachyon for their Ironwood-related work.
The migration drove activity higher. Transaction volume grew 21% to about 710,000; more than half of non-coinbase transactions touched the shielded pool, up from about 27% in the second quarter. Network difficulty rose 55% on average. Network REV, the real economic value of fees paid by users, grew about 85% to about $204,000.
Institutional on-ramps widened in both directions. Grayscale's ZCSH - the first U.S. exchange-traded product offering spot ZEC exposure - began trading at the end of August, initially holding about 387,000 ZEC, and attracted about $263 million in net inflows by quarter-end. Nasdaq-listed Cypherpunk Technologies, which owns a ZEC treasury vehicle, increased its holdings to about 326,000 ZEC and began mining.
Zakura and the Scaling Path
Delivered
Zakura is a consensus-compatible Zcash full node built on the Zebra codebase. Version 1.0.0 launched on July 15, ahead of the July 18 end-of-life for the original node client zcashd. It ships with Ironwood support, a compatibility mode that emulates the zcashd interface for existing integrations, and snapshot-based startup. Zakura nodes began mining mainnet blocks on launch day.
The first version sharply reduced sync time. Block sync that took 20 hours 46 minutes on Zebra took only 4 hours 20 minutes on Zakura. Starting from an approximately 11GB pruned snapshot took only 1 minute 50 seconds, which the team calculated to be 680 times faster than syncing over the standard peer-to-peer network. Version 1.2.0 cut block processing time by 87%. Version 1.5.0 at the end of September brought miners faster block submission and more stable templates.
Cryptography Layer
Released on August 29, Zakura Common is an open-source cryptography optimization toolkit that can be adopted by Zcash wallets and nodes. On Zakura's benchmark hardware, mobile proof generation sped up more than 14x, desktop proof generation more than 5x, Sinsemilla hashing more than 21x, and zk-SNARK verification 4 to 8x. For users, proof generation time for a standard two-input, two-output shielded transaction on an iPhone fell from more than 3 seconds to about 200 milliseconds. Zakura 1.3.0 migrated to the new stack on August 31, with the Vizor wallet among the earliest adopters.
The Route to the Throughput Target
The team set a target of about 50,000 private transactions per second. When Zakura launched, Ojha estimated the network's shielded capacity at about 2.92 TPS under 75-second blocks and placed the target two to three years out. With Zcash's existing cryptography, that load requires nodes to process more than 500MB of data per second. Every shielded transaction carries its own proof. Zakura handles node performance and networking. Project Tachyon's recursive proofs let nodes replace verifying thousands of proofs with verifying one aggregated proof, reducing the requirement to about 100MB per second; Valar Group's private information retrieval targets the wallet side, where the current stack's throughput ceiling is close to 1 TPS.
Tachyon began integrating into Zakura Common on September 28. Zakura also stress-tested a large distributed node network and reported that Zcash is already capable of supporting 25-second blocks - NU7 currently plans to implement this change in November. Bowe co-invented the Halo 2 proof system and has been responsible for Zcash's zk-SNARK implementation from launch through every network upgrade since. Both teams initially got started with private donations from ZEC holders. In September, holders voted to pay for their Ironwood work from the protocol holder-controlled fund.
Network Performance: Transactions Rebound, Difficulty Rises Faster
Zcash processed 710,047 transactions in the third quarter, up 21% from 585,178 in the second quarter and triple the 228,313 in the third quarter of 2025. Over the past five quarters, only the 933,284 in the fourth quarter of 2025 was higher. Average daily transactions were about 7,700, versus about 6,400 in the previous quarter; momentum built through the quarter: July averaged about 4,400 per day, rising to about 13,300 in September, nearly tripling. Some of the growth came from technical factors. The wallet-led Orchard migration itself generated traffic, and Zodl's privacy-preserving migration split balances into standardized denominations (0.1, 0.5, and 1 ZEC), sent at randomized times, by design producing multiple transactions per migrated balance.
Hashrate rose faster than activity. As the closest proxy for hashrate, network difficulty averaged 222.6 million in the third quarter, 55% higher than the second-quarter average of 143.4 million. Since the third quarter of 2025, difficulty has risen every quarter, more than tripling from 66.3 million. Block production remained on schedule. Blocks averaged one every 75.3 seconds, with 105,487 produced against a target of 75 seconds.
The median transaction fee in the third quarter averaged about $0.13, up from about $0.07 in the second quarter; the average fee rose from $0.18 to $0.24. Zebra also supports a lower conventional fee, 1,000 zatoshi per logical operation, down from 5,000, allowing wallets to choose to charge lower fees.Zcashd reached end of life on July 18, with its operators moving to Zebra and Zakura. Zebra 6.0.0 delivered full Ironwood support ahead of activation; Zakura 1.0.1 fixed an inherited vulnerability in a release coordinated with the Zcash Foundation, with no user funds at risk. Ironwood itself activated on July 28 at block 3,428,143, a week later than originally planned, after exchanges and wallet providers requested more preparation time. The upgrade introduced the v6 transaction format for the new pool.
Supply and Usage
Shielded supply rebounds
In Q3, the total ZEC in shielded pools rose 11.8%, from 4.42 million to 4.94 million; the shielded share of supply rose from 26.3% to 29.1%. The quarter recovered most of the outflow following the Q2 Orchard incident. The transparent pool saw a net quarterly decline of 378,000 ZEC, while shielded pools added 523,000.
The composition of shielded supply changed completely. Ironwood started from zero at activation on July 28 and held 4.06 million ZEC at quarter-end. More than 1 million ZEC migrated within seven days of activation, and more than 3 million by August 16. Orchard held about 3.66 million ZEC when Ironwood sealed it and stopped new deposits, then fell to 379,000. Sapling declined 23% to 479,000 ZEC, and Sprout held about 22,000.
Usage shifts to the new pool
Shielded activity grew faster than the network as a whole. In Q3, transactions touching shielded pools accounted for 52.7% of all non-coinbase transactions, up from 27.4% in Q2. Ironwood alone recorded 226,452 transactions after activation, plus another 24,976 Orchard-to-Ironwood mixed transactions carrying migrated value through the turnstile. By transaction count, withdrawals from Ironwood to transparent addresses exceeded deposits, at 159,020 versus 67,429, though net value still flowed into shielded pools.
By August 10, Zodl made migration a guided wallet feature across all supported platforms; its "Migrate with Privacy" mode implements ZIP-318, so cross-pool transfers do not expose exact balances. Hardware support followed. Vizor enabled shielded ZEC for Ledger devices in mid-September, and Ledger Wallet added shielded Zcash support on September 23.
Total transfer value rose 17% to 87 million ZEC. Of that, Ironwood carried 12.4 million ZEC in its first two months, Orchard-to-Ironwood mixed transfers carried another 3.3 million, and transparent transfers accounted for most of the remainder.
At quarter-end, the anonymity set of the Sapling and Orchard pools was 124.5 million note commitments, versus 124.1 million at the end of Q2. The metric predates Ironwood and does not yet count Ironwood notes, so it understates the privacy set available to migrated users.
Market Structure and Economics
Fees and network REV
In Q3, network REV—the real economic value of fees paid by users—was about $204,000, up 85% from about $110,000 in Q2. The fee mix changed with migration. Transparent transactions contributed 60.1% of fee value in Q3, down from 83.6% previously; Ironwood contributed 25.9% in its first two months, and Orchard-to-Ironwood mixed transactions contributed another 1.7%.
Starting in November, NU7 will change how these fees work. The network sustainability mechanism moves 60% of transaction fees out of circulation, then recycles that ZEC back to miners through future block rewards, without changing the 21 million maximum supply. Holders voted with 96.6% support to delay reissuance to February 2031, so fees will accumulate for several years before returning to the security budget.
Institutional on-ramps
Grayscale converted its Zcash Trust into ZCSH, the first U.S. exchange-traded product offering spot ZEC exposure, which began trading on August 25. The fund initially held about 387,000 ZEC, custodied by Coinbase, with Jane Street and Virtu as authorized participants. It recorded about $263 million in net inflows by quarter-end, with about $996 million in assets under management as of September 25. The fund's early fee revenue is designated for Zcash ecosystem development. On September 25, Grayscale filed for a second product—an options-based ZCSH High Income ETF, which will not directly hold ZEC.
Nasdaq-listed Cypherpunk Technologies, which has a ZEC treasury vehicle, increased its holdings from 314,186 ZEC to 326,417 ZEC during the quarter, worth about $463 million at quarter-end. The increase included 3,023 ZEC produced by its own mining operation, which launched in mid-August. Cypherpunk also appointed Valar Group founder Dev Ojha as an advisor.
Liquidity and access venues
ZEC spot trading volume on centralized exchanges rose 26% to $37.2 billion across about 13 venues.
On BNB Chain, ZEC circulates as a Binance-Peg token, with supply rising in a one-time jump from 120,000 to 220,000 on September 7, its first change since March 2024. On NEAR—home to the NEAR Intents Swap service used by Zodl and Vizor for in-wallet ZEC swaps—bridged ZEC fluctuated between about 25,000 and 39,500, hitting a low on September 16; average daily ZEC volume through the service rose sixfold that week. It ended the quarter at 29,300, down from 37,400 at the end of Q2. The wrapped ZEC asset cbZEC launched on Base on September 1, bringing ZEC as collateral into EVM DeFi.
Mining and block rewards
In Q3, block subsidies paid 141,882 ZEC. Miners received 128,939 ZEC of that, and the Zcash Community Grants funding stream received 12,893 ZEC; the development fund reserve accumulated issuance shares outside those two flows, rising from about 46,900 ZEC to 66,700 ZEC. Per-block issuance is fixed, so these totals track block production, which held at 75.3 seconds per block.
Mining capacity continued to grow. Average difficulty was 55% higher than in Q2, and the daily average number of distinct reward-receiving addresses was 16.8, versus 16.2 in Q2. That figure measures payout addresses, not independent operators. With difficulty climbing while the address count stayed stable, the added hashrate did not arrive through more concentrated payout addresses.
A publicly listed miner joined the network. Cypherpunk Technologies began mining ZEC in mid-August through its subsidiary Cypherpunk Mining, producing 3,023 ZEC from August 18 to August 31, all of which went to the treasury. It then appointed Amanda Fabiano to its board to support expansion. Fabiano previously held senior mining and investment roles at Fidelity and Galaxy.Mining software also improved on both node clients. Zakura 1.4.0 adds fast proofs for shielded coinbase block templates; Zakura 1.5.0 in late September brings faster block submission and more stable templates. Zebra 6.4.1 adds support for end-of-support shutdown around November 2, before NU7; the emergency security release 6.4.2 on September 28 recalls 6.4.0 and 6.4.1, so Zebra operators need to upgrade before NU7.
NU7 changes block times, but the issuance path is unchanged. Blocks will go from 75 seconds to 25 seconds; under ZIP-218, the per-block subsidy is divided by 3 again, keeping total issuance per unit of time unchanged. The halving is preserved. Holders preserved the fixed-interval reduction of new issuance over a smoother curve with 98.9% support; fees collected by the NSM flow back to miners through block rewards starting in February 2031, supplementing the security budget as the subsidy continues to halve.
Organizational and Technical Progress
Ironwood and Its Proof
Ironwood closes the Orchard chapter that opened in the second quarter. The Orchard defect allowed counterfeiting within that pool, but the turnstile prevented any such value from entering the total supply; developers found no evidence that it was ever exploited. Ironwood replaces Orchard with a new pool and makes Orchard exit-only, with every withdrawal passing through the turnstile. The Zakura team framed the migration around three goals: a trustworthy and reliable shielded pool, proving that no counterfeiting occurred in Orchard, and rebuilding circulating supply.
The new pool comes with a machine-verified proof written in Lean, containing more than 2,700 theorems, completed over more than a month by three teams led by Tal Derei (Project Tachyon), Daira-Emma Hopwood (ZODL), and Gregor Mitscha-Baude (zkSecurity). It formally verifies that, under the stated design assumptions, Ironwood cannot create undetectable counterfeit ZEC. The proof covers only Ironwood and cannot make the same claim about Orchard history. The Zakura team calls Ironwood the first shielded pool formally verified against a supply reliability vulnerability. Independent audits commissioned by the Valar and Tachyon teams found no major issues.
Ironwood also provides quantum recoverability under ZIP 2005. Future quantum computers cannot retroactively steal or counterfeit funds in the pool, and holders retain a path to recover their ZEC. The design does not make today's transactions quantum-safe; a separate recovery protocol still needs to be designed and activated.
Two Holder Votes
The NU7 vote balances broad participation and ballot privacy. Only ZEC spendable in Ironwood at the August 24 snapshot could vote. Ballots were encrypted, and each ballot was split into 16 unlinkable fragments, so verifiers could count totals without knowing who voted or how much ZEC they held. Valar Group built the protocol using homomorphic encryption and private information retrieval. Voting ran from August 25 to September 14.
Of about 3.6 million eligible ZEC, nearly 2.4 million participated, far above the 1 million ZEC voting threshold and far above the 7.25% of circulating ZEC that participated in the February sentiment survey. Holders supported 25-second blocks (99.9%), preserving the halving (98.9%), choosing to deliver NU7 as soon as possible with a September 30 readiness deadline (99.3%), and delaying NSM reissuance to February 2031 (96.6%). They also voted to retire legacy transactions from the Sprout pool.
The second vote ended on September 29 and put the protocol holder-controlled development fund to use. The fund receives 12% of block rewards, and each grant requires at least 420,000 ZEC in votes and simple majority support. Holders approved 17 of 37 retroactive grant proposals, worth about $8.39 million, or 93.1% of the $9.01 million requested; the proposal with the highest turnout reached 2.18 million ZEC. They granted $1.203 million to Valar Group for Ironwood development and reimbursed $599,000 in external audits it paid; granted $1 million to Project Tachyon for Ironwood zk-SNARK formal verification ($738,942 requested, plus a separately nominated $261,058 bonus); and granted $1.95 million to ZODL for core protocol work in the first half of the year. Every Valar and Tachyon grant passed with more than 99.5% of the vote.
How the Ecosystem Collaborated
Ironwood was jointly proposed by ZODL, Project Tachyon, Valar Group, the Zcash Foundation, and Shielded Labs. These five organizations agreed on the final voting language and, after the results were in, unanimously agreed on the content and timeline of NU7. The Zcash Foundation publicly welcomed Zakura as a full node built on its Zebra codebase, and the two teams also coordinated security releases.
Forward Roadmap
NU7 is the next delivery, and the timeline is set. The upgrade activated on the public testnet at block 4,465,026 on October 4, earlier than the previous estimate of about October 6. Developers plan to set the mainnet activation height on October 20 after reviewing testnet results. The mainnet activation target is November 5. Zakura provided a live NU7 dashboard and a testnet faucet for operators testing the upgrade.
The upgrade contents follow the holder votes:
25-second blocks (ZIP-218), three times faster than the current rate, with the per-block subsidy reduced by the same proportion so that issuance per unit of time is unchanged.
Network Sustainability Mechanism (ZIP 235 and ZIP 237), moving 60% of fees out of circulation and reissuing them through block rewards starting in February 2031, while preserving the halving schedule.
Retiring v4 transactions, ending spending from the legacy Sprout pool.
NU7 does not introduce a new transaction format and should not have a major impact on wallets, but full nodes, indexers, and block explorers need to upgrade. Zcash Shielded Assets is still under discussion and is not part of NU7.
Scaling work will continue after NU7. Project Tachyon began a series of code releases, specifications, ZIPs, and audit reports on September 28, while starting integration into Zakura Common. Tachyon's shielded pool is built around proof aggregation and caps on the state validators must store, and is the next protocol step toward throughput goals. Valar Group's private information retrieval work addresses wallet-side constraints, and Zakura's experimental peer-to-peer transport targets worst-case block propagation below 500 milliseconds.
The post-quantum protection plan is a phased, multi-upgrade strategy. Ironwood's quantum-recoverable notes are the first step; before cryptographically relevant quantum computers arrive, vulnerable pools will be disabled while holders can still recover funds.
Zcon7 will be held in Cancun from October 27 to 29, coinciding with Zcash's tenth anniversary on October 28, one week before NU7 reaches mainnet.
Conclusion
In the third quarter of 2026, Zcash moved from recovery to the first phase of scaling. The quarter began with the Orchard pool still carrying the aftermath of the second-quarter incident; it ended with 4.06 million ZEC in a formally verified replacement pool and the shielded supply share back up to 29.1%. Holders voted twice: once to determine the contents of the next upgrade, with nearly two-thirds of eligible supply participating; and once to pay $8.39 million from the holder-controlled fund for completed protocol work.Zakura is the quarter's defining result. It cut full sync from most of a day to about four hours, snapshot startup to under two minutes, and sped up shielded transactions in mobile builds by more than 14x. It also pushed performance work into shipped software, reaching users without a network upgrade. As of Q3, the first layer toward the 50,000 TPS target is live, the second layer (Tachyon recursive proofs) has begun integration, and the next protocol-level change—25-second blocks—has a date and has been running on testnet since October 4.
Network economics moved in the same direction. Hashrate rose, difficulty grew 55%; network REV grew 85% to about $204K; more than half of non-coinbase transactions used the shielded pool. Institutional access widened through Grayscale's ZCSH, which drew about $263M in net inflows by quarter's end; Cypherpunk's treasury and mining operations also grew.
The open item carried over from Q2 is Orchard itself. About 379K ZEC remains in the pool, and every coin that leaves passes through the turnstile, a migration that builds for Orchard's history exactly the evidence a formal proof cannot provide.
Three measurable milestones stand in Q4: NU7 reaching mainnet on November 5, bringing 25-second blocks and network sustainability mechanisms; the Ironwood migration clearing Orchard's remaining 379K ZEC; and Tachyon moving from integration to deployment.