Tezos Launches Quantumnet to Test Post-Quantum Cryptography
The first significant deployments of quantum-resistant technology are emerging not in laboratories or defense contracts, but in public ledgers. Here, every posted signature remains permanently readable, years before any machine can forge one. Tezos has initiated testing of this upgrade on its live infrastructure with a public testnet named Quantumnet.
Inside Quantumnet’s Operations
On September 24, 2026, engineering teams from Nomadic Labs, Trilitech, and other contributors launched Quantumnet. This experimental public testnet is the first to implement post-quantum consensus and cryptographic code within the Tezos protocol. It is not production software; however, anyone, including validators, can run it. The teams are seeking feedback before drafting a protocol amendment for the mainnet.
This effort can be likened to replacing every door lock in a city while the machine capable of picking the old locks is still in its manufacturing phase. The new locks are heavier and more costly to install, impacting every doorframe. Quantumnet aims to quantify these new costs.
The testnet replaces several instances where the core protocol relies on quantum-vulnerable cryptography, according to the Tezos Spotlight announcement:
- Transaction signatures are being migrated to ML-DSA-44, the NIST-standardized lattice scheme, building upon the tz5 account type introduced in the June 2026 Ushuaia upgrade.
- Consensus attestation aggregation now uses a post-quantum proof system from LeanEthereum, which can combine over 1,400 signatures per second on standard hardware.
- The Data Availability Layer has adopted ZODA, a scheme the team states adds virtually no communication overhead.
- Randomness generation completely eliminates the Verifiable Delay Function, replacing it with a weighted round-robin system that requires no randomness.
The Impact of Larger Signatures
This presents an engineering challenge, not merely a configuration change. Elliptic-curve signatures, such as Ed25519, are compact—64 bytes for a signature and 32 for a public key. In contrast, post-quantum schemes are not. ML-DSA-44, standardized as FIPS 204, generates signatures around 2,420 bytes with public keys close to 1,312 bytes. This represents a two-order-of-magnitude increase in the data associated with each transaction.
On a blockchain, this data is permanent; it is stored, replicated across validators, and verified in every block, indefinitely. Larger signatures lead to greater storage growth per block, increased bandwidth between nodes, and more extensive verification work. Quantumnet’s purpose is to determine the actual cumulative cost of these factors under load.
Currently, no deployed measurements, regarding signature overhead, verification latency, or throughput changes, exist in any accessible source. This benchmarking is the immediate task, and the parameter sizes mentioned above are derived from the NIST standard, not from Tezos’ own measurements.
Why a Ledger Is Considered Long-Lived Data
The inherent threat model here is “harvest-now-decrypt-later.” An adversary could record public signatures today and forge them once a sufficiently capable quantum computer becomes available. Most systems can rotate keys before such a threat materializes.
However, a public ledger cannot retract previously published information. Every signature ever recorded on Tezos remains a persistent target, making a blockchain one of the few systems that must migrate before the threat becomes real, rather than after.
This perspective also explains the institutional urgency surrounding long-lived data and compliance timelines. Google’s 2029 deadline for migrating its own systems to post-quantum cryptography illustrates the level of risk perceived by major operators.
Timeline for Implementation
How much time remains? Vendor roadmaps provide a general indication. QuEra and HPE recently collaborated on fault-tolerant machines for high-performance computing, with QuEra aiming for a system in the 256-plus logical qubit range to be cloud-available around 2028.
These are vendor projections. No cryptographically relevant advantage against elliptic-curve signatures has yet been demonstrated. The rationale for testing now is that migration on live consensus infrastructure takes years, and the timelines for the threat and the solution are converging.
Quantumnet’s Capabilities and Limitations
Quantumnet demonstrates the operational viability of these schemes and the protocol’s adaptability. However, it does not yet prove mainnet performance, as testnet conditions typically involve lighter loads and, as Tezos notes, higher hardware requirements.
It is important to distinguish that ML-DSA-44 and the hash-based options belong to the NIST-standardized family (FIPS 203, 204, and 205), not a proprietary construction. Furthermore, QKD or QRNG address entirely different problems; neither is a substitute for the signatures Quantumnet is upgrading.