Contents
Post-quantum address proposal bc1z
Developer Zaenus published a Bitcoin Improvement Proposal draft on November 13, 2026 2. The proposal introduces a quantum-resistant address format starting with the prefix bc1z 2. This format incorporates the ML-DSA-65 post-quantum signature algorithm specified in FIPS 204 2. Zaenus created the proposal within the safe-btc project repository on GitHub 2. The design aims to protect funds against potential decryption by Shor's algorithm running on a cryptographically relevant quantum computer 2. The author estimated that approximately 6.9 million BTC, representing about $586 billion, resides in addresses with exposed public keys 2. Once a public key appears on the public ledger, a quantum computer can derive the corresponding private key to steal the funds 2. To prevent this vulnerability, the proposed address format hides the public key inside a SHA-256 commitment until spending 2.
The bc1z design uses a 32-byte hash commitment to conceal the underlying public key 2. The address payload consists of the human-readable string bc1z, witness version 0x02, and the 32-byte SHA-256 hash 2. The public key remains hidden until the transaction owner creates a spend witness 2. This mechanism functions similarly to existing Pay-to-Witness-Public-Key-Hash outputs 2. Zaenus released a Python reference implementation containing utilities for key generation, signing, signature verification, UTXO migration, and quantum canary monitoring 2.
Witness requirements and soft fork design
Spending a bc1z output requires a specialized witness structure in the transaction 2. The witness payload includes an ML-DSA signature of 3,309 bytes 2. The witness payload also includes the original ML-DSA public key of 1,952 bytes 2. Combining the signature and public key yields a total witness size of 5,261 bytes per input 2. The proposal relies on a soft fork mechanism for network deployment 2. Legacy nodes interpret outputs with witness version 0x02 as anyone-can-spend transactions 2. Upgraded nodes validate that the SHA-256 hash of the 1,952-byte public key matches the address commitment 2. Upgraded nodes also verify the 3,309-byte ML-DSA signature against the revealed public key 2.
Signature size overhead and SNARK compression
Post-quantum signatures require significantly more data space than ECDSA or Schnorr signatures 2. Ten individual ML-DSA signatures require approximately 33 KB of transaction witness space 2. To reduce this data load, Zaenus developed a signature aggregation prototype 2. The prototype wraps ML-DSA signatures inside a zero-knowledge SNARK proof 2. The SNARK wrapper compresses ten individual signatures from roughly 33 KB down to about 2 KB 2. Zaenus requested feedback from developers regarding ML-DSA-65 parameter trade-offs, address design choices, soft fork deployment strategies, and UTXO migration paths 2.
STARK proofs for UTXO verification
A separate research initiative on Delving Bitcoin presented a STARK proof system for the Bitcoin UTXO set on November 13, 2026 1. The developer released a 1.6 MB proof seal that covers chain state up to block 972,500 1. A standard web browser verifies the 1.6 MB seal file in roughly 0.5 seconds 1. The proof verifies header rules from genesis to the chain tip, including SHA256d linkage, proof of work, retargeting, median time past, minimum version, and total chainwork 1. It excludes only the wall-clock time-too-new check 1. The system replays every block from block 935,001 to block 972,500 1. Replay verification covers raw block data, Merkle roots, txid values, wtxid values, transaction structural checks, amounts, and coin creation 1.
The UTXO set output from block 972,500 matches the serialized hash produced by Bitcoin Core 1. Bitcoin Core dumptxoutset commands at block 972,500 produced the hash 9f92d143...c8c4eb 1. The STARK proof commits to a Merkle root over the UTXO set 1. Users can generate inclusion and non-inclusion proofs of roughly 1 KB per outpoint 1. Single lookups download approximately 2 MB segments of the UTXO set 1. The current proof relies on Core assumeUTXO snapshot trust at block 935,000 1. The proof does not check individual scripts or signatures, matching the trust model of assumevalid 1.
Prover performance and Lean circuit verification
The prover stack uses Stwo circle STARKs over the M31 field through a Zig port and a CUDA backend 1. A large block containing 2^23 trace rows proves in 2.7 seconds on an Nvidia H100 GPU 1. Circuits written in Lean 4 define the machine-checked constraints for header rules, SHA-256 execution under FIPS 180-4, block checks, and set transitions 1. Lean specs underwent differential fuzzing against Bitcoin Core C++ code 1. Testing evaluated 750,000 header-rule cases and 100,000 block and transaction cases with 0 mismatches 1.
Running chainproof verify-snapshot on a Bitcoin Core version 31.1 mainnet node at block 972,500 validated a snapshot with 164,539,066 coins stored in a 9.39 GB file 1. Proving the complete block history from genesis is estimated to require about 325 H100-hours 1. Generating this full genesis proof carries an estimated cost of $1,000 to $1,500 USD 1. Trusted elements in the current system include Stwo, FRI, LogUp, and Fiat-Shamir in the random oracle model 1.
What is not yet established
- Whether developer consensus will support activating the bc1z soft fork prior to the existence of a cryptographically relevant quantum computer.
- The final parameters for ML-DSA signature compression trade-offs between proof size and computation time.
- External security review status for the Lean circuit specifications and the Stwo prover implementation.
Frequently asked questions
What is the bc1z address format?
It is a proposed quantum-resistant Bitcoin address format using witness version 0x02 and ML-DSA-65 signatures 2.
How large is the spend witness for a bc1z address?
Spending a single bc1z input requires 5,261 bytes, comprising a 3,309-byte signature and a 1,952-byte public key 2.
Sources
- Delving Bitcoin — A STARK proof of Bitcoin's UTXO set at block 972,500, with circuits proven in Lean (2026-11-13) https://delvingbitcoin.org/t/a-stark-proof-of-bitcoins-utxo-set-at-block-972-500-with-circuits-proven-in-lean/3070
- Delving Bitcoin — BIP Draft Post-Quantum Address Format (bc1z) using ML-DSA-65 (2026-11-13) https://delvingbitcoin.org/t/bip-draft-post-quantum-address-format-bc1z-using-ml-dsa-65/3067
Revision history
Published