KIP-21 Merged: The Partitioned Sequencing Commitment Lands in Rusty Kaspa
April 20, 2026
Kaspa developer Max Biryukov's KIP-21 implementation has been approved and merged into rusty-kaspa, marking the final major milestone before the Toccata hard fork feature freeze and TN12 testnet reset. Pull Request #943, spanning 132 commits, introduces the partitioned sequencing commitment — the architectural change that makes zero-knowledge applications economically viable on Kaspa's 10 BPS proof-of-work network.
What KIP-21 Changes
Kaspa already commits to transaction sequencing via a single recursive hash (KIP-14, KIP-15). The problem: every L2, subnet, and zk application has to prove against the entire DAG's activity. Proving costs scale with total network throughput, not with the application itself.
KIP-21 partitions the commitment into lanes. Each L2, subnet, or zk app gets its own lane tracked in a 256-bit Sparse Merkle Tree. Applications prove against their lane only. Proving costs now scale with application usage, not network load.
This is the unlock for zk on Kaspa.
What PR #943 Ships
no_std 256-bit Sparse Merkle Tree with bitmap-compressed proofs for ZK circuits
Versioned, fork-aware storage layer with incremental O(256) updates
Full virtual-processor integration with atomic flush alongside UTXO state
P2P protocol v9 with streaming SMT state sync during IBD
Eight new BLAKE3 domain-separated hashers for lane structure
Tightened subnet ID validation rules
Michael Sutton's Approval
"LGTM. This is a seriously impressive piece of work. Complicated and powerful, but the code remains sound and well designed throughout — elegant, beautiful work."
What's Next on the Toccata Roadmap
Toccata mainnet activation is scheduled for a window between June 5–20, 2026. The original May 5 target was pushed specifically so the sequencing architecture could land correctly. KIP-21 is the piece that needed to be right.
Remaining steps: feature freeze, TN12 reset, master branch merge, auditing, TN10 transition rehearsal, and mainnet activation date hardcoding.
Why It Matters
The sequencing commitment is the interface between Kaspa L1 and every rollup, subnet, zk application, and bridge that wants to build on top of it. Partitioning it per-application is what makes those systems practical at Kaspa's throughput.
One component. Merged. The implications play out over the next cycle.