Author: Coindesk
Translated by: TechFlow DeepTide
DeepTide Guide: Zcash currently only processes about 1 privacy transaction per second. The goal of the new node Zakura is to achieve 50,000 TPS while maintaining full privacy—the processing scale of Visa and Mastercard. This is a significant signal for everyone focusing on payment privacy and blockchain scaling: privacy and performance are no longer a trade-off.
Those rebuilding Zcash have a dream: to reach the level of global payment giants Visa and Mastercard, processing tens of thousands of payments per second, while maintaining full verifiability and strong privacy guarantees.
The first piece of this plan is Zakura, a new full node software whose version 1.0.0 was released on Wednesday. It is maintained by Sean Bowe, a founding member of Zcash's zero-knowledge cryptography, and Dev Ojha, co-founder of Osmosis and currently in charge at Valar Group. Both teams are funded by private ZEC donations, not by a company or foundation.
"Our dream is to support global payments. Mastercard and Visa process over 50,000 transactions per second; this is our baseline. Under Zcash's current cryptographic scheme, that volume requires nodes to process over 500 MB per second of throughput," a blog post stated. "The existing tech stack doesn't meet this goal. The cryptographic technology our team is developing will close most of that gap."
A full node is a program that keeps a complete copy of the blockchain, here referring to the Zcash ledger, and independently checks whether every transaction complies with the network's rules. Zakura is a fork of the Zcash Foundation's node software, Zebra—meaning it started from the foundation's official code and was rebuilt.
Consensus rules are the shared rulebook executed by every node, determining which blocks and transactions the entire network accepts as valid. If a node applies different rules, it forks and stops following the same chain as the other nodes.
Pruning, Snapshots, and Compatibility
Zakura can also prune, meaning it deletes old blockchain data that the node no longer needs, significantly reducing disk usage. This compresses the chain small enough that the team released a ready-made copy of about 11 GB after stripping old data. New nodes can download this directly instead of pulling the entire history block by block from other nodes.
This allows a node to go from zero to running in under two minutes, which the team calls "680 times faster."
A compatibility mode further replicates the interface of the original client, zcashd, which is reaching its end-of-life on July 18. This means wallets and exchange integrations built for it can continue to work as usual.
Throughput Goals and Tachyon's Role
The reason for building all this is an arithmetic problem.
Mastercard and Visa process over 50,000 transactions per second, a number the team calls a "baseline, not a target." Zcash's current cryptographic scheme requires nodes to receive and verify over 500 MB of data per second to keep up because each privacy transaction carries a proof, and that proof is large.
This is roughly equivalent to receiving a full DVD's worth of data continuously every 10 seconds, a rate no current Zcash software can approach. But the key is understanding why each bottleneck exists.
Bowe's Tachyon project is addressing this problem with recursive proofs, where one proof can attest to the validity of thousands of other proofs, drastically reducing the amount of data that must be checked at the consensus layer.
Under the Tachyon scheme, nodes validate a single proof instead of thousands. The team claims this will reduce consensus data needs from 100 MB/sec to 500 MB/sec, which they believe is technically achievable with careful engineering.
Wallet Bottleneck and Valar's PIR Solution
Wallets face a different problem. Because Zcash hides the recipient of a transaction, a wallet cannot ask a server which transactions belong to it without revealing itself. It must pull everything and test each one, which is why wallet software tops out at about 1 transaction per second.
To eliminate this bottleneck, Valar Group is researching Private Information Retrieval (PIR) technology, which allows a wallet to fetch its own data from a server without the server knowing which entries were requested.
Fast Block Propagation
Fast block propagation means broadcasting newly mined blocks across the blockchain network as quickly as possible. Zakura is the software layer responsible for this task.
It must be fast enough to transmit new blocks between nodes to enable the heavy proof and wallet traffic. It comes with an experimental system, currently disabled by default, designed to deliver each block to every node within half a second.
Near-term testing of these ideas will arrive in late July. Ironwood, formally named NU6.3, will activate at mainnet block 3,428,143, approximately 8:00 AM ET on July 28. Zakura has supported it since its release.
Bowe stated on July 10 that all major organizations have committed to activating at this height, a week later than originally planned because exchanges and wallet providers requested preparation time.
The Origin of Ironwood
Ironwood exists because of a bug that nearly destroyed Zcash in June. The so-called shielded pool is the private side of the network, where amounts and participants are hidden, with zero-knowledge proofs serving as evidence of mathematical operations.
On May 29, Shielded Labs researcher Taylor Hornby discovered a soundness bug in the proof circuit of Orchard (the latest shielded pool) that allowed an attacker to mint fake ZEC on-chain without a trace. This bug had existed since Orchard's activation in May 2022.
Developers disabled Orchard through an emergency response completed on June 2, then restored it on June 3 with a corrected circuit via the NU6.2 hard fork at block 3,364,600.
The patch couldn't address the four years the bug was open. Zero-knowledge proofs reveal nothing beyond the fact that verification passed, so there is no on-chain record of what any Orchard transaction moved. No one can prove that fake ZEC was never created.
Ironwood was built to solve this problem. So-called "turnstiles" at the pool boundaries limit what can leave and enter, leveraging the fact that the amount of ZEC entering or exiting the shielded pool is public, even if transactions inside the pool are private. By enclosing Orchard to block new deposits, the turnstile becomes the only exit, trapping any fake coins inside.
In simple terms, honest balances can migrate out over time, while fake coins may be prevented from exiting entirely and entering the circulating supply. This setup traps any attempt at excess supply at the boundary, restoring confidence in the token supply.
