KZG polynomial commitments let a prover commit to a polynomial and prove a claimed evaluation with a small cryptographic witness. The name comes from Aniket Kate, Gregory M. Zaverucha and Ian Goldberg, whose ASIACRYPT 2010 work is titled Constant-Size Commitments to Polynomials and Their Applications. Their extended Waterloo report, Polynomial Commitments, is dated 1 December 2010. This page reviews the construction and separates it from later Ethereum applications.
| Authors | Aniket Kate; Gregory M. Zaverucha; Ian Goldberg |
|---|---|
| Year | 2010 |
| Document reviewed | Polynomial Commitments, CACR 2010-10, extended report |
| Research area | Polynomial commitments and pairing-based cryptography |
| License | No explicit public-domain or reuse grant identified in the reviewed report. Ethereum specifications and implementations have separate licenses. |
What a Commitment Proves
A commitment is a compact reference to a polynomial. An opening lets a verifier check a statement such as f(z) = y against that reference. Binding means an adversary should not be able to produce conflicting accepted openings under the scheme’s security assumptions. It does not prove that the polynomial describes a correct transaction, a solvent protocol or available network data. Those require additional rules and evidence. [1, 2]
Commit, Open and Verify
Using additive elliptic-curve notation, public setup parameters include encoded powers of a secret tau. The prover combines them with polynomial coefficients to compute C = [f(tau)]₁, without learning tau. To prove f(z) = y, it constructs the quotient q(X) = (f(X) - y) / (X - z) and sends pi = [q(tau)]₁. This quotient is a polynomial when the claimed evaluation is correct. [1, 2]
A pairing check relates the commitment, value, proof and evaluation point. A single-point commitment and proof each occupy one group element in this construction. The byte length depends on the chosen encoding; Ethereum’s BLS12-381 encoding uses 48 bytes each. The 2010 paper did not prescribe Ethereum’s later blob format. [2, 3]
Small Proofs Still Have Costs
For a fixed curve and a single opening, proof length and the basic pairing check do not grow with polynomial degree. The setup’s supported degree is bounded, however, and its public parameters grow with that bound. Computing commitments and witnesses still requires work over polynomial data. Batch openings reduce witness overhead without eliminating the cost of receiving the opened values or performing other protocol checks. [1, 2]
Therefore “constant-size proof” should never be expanded into “every operation takes constant time.” Separate communication size, prover work, verifier input processing and setup storage when comparing systems.
Setup and Privacy Assumptions
If an attacker learns the setup trapdoor, it can forge evaluation witnesses. Distributed setup aims to prevent any attacker from learning it; an honest contribution must be generated, incorporated and discarded correctly. The published group elements are public—the secret scalar is what must remain unknown. [1, 2]
Binding also does not automatically provide transaction privacy. The basic deterministic commitment can be checked against guessed polynomials. The paper treats hiding separately and gives a randomized Pedersen-style construction. Ethereum blobs are public data, so their KZG commitments should not be presented as encryption. [1, 2]
Ethereum: Point Openings, Full Blobs and Sampling
Ethereum‘s EIP-4844 specifies blobs of 4,096 field elements encoded in 32 bytes each: 131,072 bytes, or 128 KiB. Blob contents are unavailable directly to EVM execution; commitments are referenced through versioned hashes and an evaluation precompile. That distinction does not remove the consensus layer’s data-availability responsibilities. [3]
In the versioned Deneb specification, verify_blob_kzg_proof takes the full blob, converts its contents to polynomial form, derives a challenge and evaluates the polynomial before checking the proof. A point-opening verifier is a different operation. The old claim that Deneb verification needs no blob data is incorrect. [4]
PeerDAS, specified separately in EIP-7594, combines erasure coding, column custody, peer requests and cell KZG proofs so nodes can check availability through subsets of data. A valid commitment alone cannot establish that missing data can be downloaded. Current C-KZG documentation distinguishes the blob-proof API from its cell-proof, recovery and sampling-related API. [5, 6]
Common Misconceptions
- “KZG is a complete zero-knowledge proof system.” It is a commitment primitive; a full proof system adds its own constraints and security argument.
- “Every polynomial commitment uses KZG.” Other constructions exist; a system’s specific backend must be checked.
- “A small commitment guarantees availability or permanent storage.” Integrity checks, availability mechanisms and retention policies answer different questions.
Social Media Sentiment
No current social-media sample was collected for this technical research reference. The construction is assessed through its stated assumptions and implementation specifications; no market or community sentiment claim is made.
Last updated: 2026-10
Related Terms
- Ethereum — a later application of KZG commitments.
- Data Availability — whether required data can be obtained.
See Also
- BLS Signatures — another pairing-based construction, with different proof obligations.
- Bulletproofs — a separate proof-system research reference.
Sources
- Kate, Zaverucha and Goldberg, Polynomial Commitments, CACR 2010-10, 1 December 2010. Extended ASIACRYPT report; definitions, constructions, setup, hiding and batch-opening qualifications.
- Dankrad Feist, KZG polynomial commitments, 16 June 2020. Researcher’s technical explanation of quotient witnesses, pairings, trapdoor risk and deterministic-commitment privacy limits.
- Ethereum, EIP-4844: Shard Blob Transactions, created 25 February 2022; Final specification accessed 8 October 2026. Blob encoding, commitments, versioned hashes and consensus-layer responsibilities.
- Ethereum consensus specifications, Deneb Polynomial Commitments, version 1.4.0. Versioned source distinguishing point verification from full-blob verification.
- Ethereum, EIP-7594: PeerDAS, created 12 January 2024; Final specification accessed 8 October 2026. Erasure coding, column sampling and cell-proof verification.
- Ethereum, C-KZG-4844 implementation, accessed 8 October 2026. Primary implementation documentation separating blob, cell-proof and recovery APIs.