Paper 2026/1477

PERSEPHONE: Zero-Knowledge Multiplicative Non-Negative Proof for Sequential Private Range Verification

Abstract

Privacy is a fundamental requirement of modern digital payment systems, which process highly sensitive financial data, including account balances and transaction amounts. Unauthorized disclosure of this information can compromise users’ financial privacy and expose them to fraud or other forms of financial abuse. A key challenge in this setting is to verify that a payment amount does not exceed the payer’s available balance without revealing either the payment amount or the balance. A common approach is to employ zero-knowledge range proofs (ZKRPs). However, because conventional ZKRPs require public bounds, verifying a private value against private bounds requires multiple ZKRP instances, resulting in substantial computational and communication overhead that can increase latency and lead to cascading failures at scale. Motivated by the sequential nature of transactions in real-world payment systems, we formalize the problem of Sequential Private Range Verification (SPRV) and propose a framework, Persephone, that exploits the sequential structure of linked verifications rather than treating each verification independently. Within Persephone, we design and leverage a novel zero-knowledge proof primitive, ZK-MultNNP, to solve SPRV efficiently. Experimental results demonstrate that Persephone reduces total proving time by at least 3× compared to a ZKRP-based baseline. Furthermore, in a digital payment scenario, Persephone completes all transaction verifications within the 400 ms Doherty threshold, compared to only 30% of verifications for the ZKRP-based baseline.