Today neptune-core v0.12.0 was released, and with it hardfork gamma was activated. The hardfork was motivated by a soundness issue in the underlying cryptography that was reported to the neptune-core team in a responsible disclosure on June 6, 2026.
The time between June 6 and today has been spent on going through the soundness of the cryptography again and again, until the most thorough security reviews came back clean, multiple times in a row. Especially Neptune’s underlying zk-STARK engine, Triton VM, has been heavily scrutinized, and we now believe that process is completed.
As some of you may know, soundness issues on privacy-preserving blockchains are harder to deal with than on see-through blockchains. If a software bug in Bitcoin allows an attacker to mint more coins than is intended, the offending transactions and blocks can easily be identified and rolled back or simply made invalid. Privacy-preserving blockchains like Neptune do not have this luxury. The “zk” in “zk-STARK” means “zero knowledge”, and in this context that means that the network that verifies the validity of the blocks learn nothing about the amounts transacted in the blocks. And with soundness issues, the proof does not make the cryptographic guarantees that the protocol requires. In other words, a clever attacker could theoretically have exploited the soundness issues and used it to inflate the total money supply to their own advantage. For this reason, hardfork gamma enforces two new policies:
- Roll the blockchain back to block height 40,299 (May 28, 2026) which marks the release of the AI tool that discovered the soundness bug. Meaning that people with access to this tool will not have had time to exploit Neptune before the rollback.
- Reset the lustration counter: All UTXOs (read: coins) generated prior to the activation of hardfork gamma will have to go through a so-called lustration barrier, where all UTXOs have to reveal their contained amounts (technically, they reveal the entire
Utxodata structure, which contains the total amounts). Since we know how many NPT are supposed to exist, we can simply start a counter at this value and then subtract from it whenever UTXOs generated prior to the hardfork is mined. If a new block were to make this counter negative, the network would reject this new block, and the lustration barrier (the ability to spend old UTXOs) would be closed. This gives a cryptographic guarantee that no unintentional inflation has occurred.
On a personal note, I’m glad that we’ve found a way to correct for past soundness errors in other ways than resetting the whole blockchain and paying out from a redemption fund. And I’m glad that the team agrees on the importance of providing a cryptographic guarantee that the total monetary supply has not been inflated due to these oversights. Only through these guarantees can we in good conscience make the claim that Neptune is a contender for the future of money.
These last rounds of reviews have been very thorough, and I hope this marks the last time an emergency halting of the blockchain happens; that this does not mark the end of Neptune, nor the beginning of the end, but perhaps the end of the beginning.