Nov 12 - Nov 13, 2026
This initiative adopts ML-DSA-65 signatures, adhering to FIPS 204 standards, and emphasizes enhanced security by committing to a SHA-256 hash of the public key. In this model, the full public key is revealed only during the spending process within the transaction witness—an approach that preserves privacy and keeps outputs compact, reminiscent of Pay-to-Witness-Public-Key-Hash (P2WPKH) schemes but with notable improvements. Unlike legacy systems vulnerable to quantum attacks once public keys are exposed, this design ensures that outputs remain secure even after the public key and signature are disclosed, mitigating risks posed by adversaries capable of exploiting Shor’s algorithm on a sufficiently advanced quantum computer. Historical exposure has left over a third of all bitcoin at risk due to visible public keys, underscoring the urgency for such advancements.
The structure of these new addresses is characterized by the prefix bc1r (following a revision from the initial bc1z), a witness version 3 byte, and a 32-byte tagged hash commitment incorporating a key type indicator. When spending, the witness reveals both the ML-DSA signature and the associated public key blob, facilitating future extensibility for other key types through additional soft forks without consuming further witness versions. The cryptographic design introduces distinct context tags for transaction and message signatures, aligning with best practices to separate on-chain and off-chain uses. Comprehensive test vectors validate key generation, derivation, and spend scenarios, ensuring compatibility and correctness across implementations. The reference implementation, provided in Python, now correctly generates bech32m checksums and defaults to testnet addresses, as mainnet outputs would remain anyone-can-spend until activation by a soft fork.
Discussion around the proposal centers on several axes: the specifics of the address format, the trade-offs between ML-DSA-65 and alternative parameters such as ML-DSA-44 (which offers smaller inputs at potentially reduced security), strategies for deploying the change via soft fork, and migration pathways for existing unspent transaction outputs (UTXOs). Migration is a substantial undertaking; moving all vulnerable UTXOs would demand significant block space and time, highlighting the necessity of establishing secure destinations before the emergence of quantum threats. While prototype efforts to aggregate signatures using SNARKs were ultimately deemed insecure—due to their own vulnerability to quantum attacks and requirements for trusted setups—the documentation includes a research note on what sound aggregation might require.
Community engagement is actively encouraged. Developers and researchers are invited to review the draft specification, accessible in the btc-safe repository under docs/bip/0001-pq-address-format.md, and participate in ongoing technical discussions, including those hosted on Delving Bitcoin. Additional resources for historical context and comparison include several prior post-quantum proposals (example pull requests and discussions, https://github.com/bitcoin/bips/pull/1832, https://github.com/bitcoin/bips/pull/2147, https://github.com/bitcoin/bips/pull/2176, https://github.com/bitcoin/bips/pull/2194), related discussions (https://delvingbitcoin.org/t/new-post-quantum-bitcoin-proposal-using-winternitz-lamport-auth-chains/2525, https://delvingbitcoin.org/t/p2wots-post-quantum-utxo-winternitz-signatures/2530), successful output-type BIPs (BIP-0016, BIP-0141, BIP-0341, BIP-0360), and the mature [SHRINCS BIP draft
TLDR
We’ll email you summaries of the latest discussions from high signal bitcoin sources, like bitcoin-dev, lightning-dev, and Delving Bitcoin.
We'd love to hear your feedback on this project.
Give Feedback