Where EMV stands today
EMV chip payments rely on two kinds of cryptography. The everyday work, proving to the issuer that a card is genuine when a transaction goes online, uses symmetric keys (Triple DES, and increasingly AES). Quantum computers pose no practical threat here, and EMVCo’s analysis of AES-128 is convincing.
Public-key cryptography plays a narrower but universal role. Virtually every chip card carries a certificate chain (RSA, or ECC on C-8 cards) so terminals can check it offline. Where offline enciphered PIN is used on contact cards, the terminal encrypts the PIN with the card’s public key (RSA on most cards), and the card decrypts and checks it. In Open-Loop transit implementations, gates open based on offline card authentication (RSA, or ECC on C-8). C-8, EMVCo’s newest contactless kernel, uses elliptic curve cryptography (ECC) to set up session keys and keep card data private over the air. RSA and ECC are exactly the algorithms a large quantum computer could break.
The quantum problem
The world has already set its dates. NIST’s draft transition plan deprecates RSA-2048 after 2030 (EMV’s longest keys are 1,984 bits) and disallows RSA and ECC entirely after 2035. The UK roadmap follows the same window; the EU’s is stricter, expecting high-risk systems, with finance among the examples, to migrate by 2030. Financial regulators and auditors are starting to ask about it.
Whatever the real timeline for a quantum computer, from around 2030 a payment technology built on RSA and ECC will look outdated. Issuers, regulators and the press will ask a simple question: is EMV quantum-safe? Today the honest answer is no, for both contact and contactless.
Time is also short. EMVCo itself says terminals need at least ten years to adopt a change, and counting back from 2035 puts the starting point in 2025–2026. Meanwhile C-8, the natural place for a fix, has not yet become the common kernel the industry hoped for.
What EMVCo has said so far
The EMVCo Security Working Group has published two papers. Their message is calm and well argued: there is no immediate quantum threat to EMV chip.
- 2024 paper. AES-128 is safe against quantum attack, and there is no need to move to AES-256. Breaking RSA or ECC would need a quantum computer many orders of magnitude more capable than today’s, unlikely before 2040. Public-key cryptography matters mainly for offline transactions, which are becoming less relevant.
- 2026 paper. A quantum computer able to break RSA-2048 may be decades away, or may never be built. Qubit counts are doubling roughly every five years, which implies about 60 years to reach the 900,000 qubits currently estimated. EMVCo will watch a set of milestones (“canaries”), such as a genuine quantum factoring of the number 21, before acting. “Harvest now, decrypt later” is seen as a government concern, not an EMV one, and offline transactions are under 1% in most countries.
EMVCo has announced a third paper that will set out the risk to EMV and the mitigation options. Much of the analysis so far is sound. The questions below are about what that third paper will commit to.
Questions for the third paper
We look forward to the third paper and hope it will answer five practical questions.
- Which clock sets the plan? If terminals need ten years and regulators have set 2035, does waiting for a canary mean EMV will still depend on RSA and ECC after the deadline?
- How should the industry answer the question “Is EMV quantum-safe?” Offline transactions may be under 1%, but public-key cryptography sits on virtually every card and terminal. Regulators and auditors will ask whether the payment stack uses quantum-vulnerable algorithms. They won’t ask what share of transactions depend on them. The honest answer today is yes, and it applies to the whole spec. Will EMVCo give issuers and acquirers a clear, positive story for their regulators?
- What about C-8 privacy? C-8 protects card data over the air with ECC. Admittedly, harvest-now-decrypt-later would only yield expired payment credentials. But recorded traffic could still reveal, years later, which card – and so which person – paid where and when. Does that also put an expiry date on one of C-8’s main features?
- How will the hardest case be solved? Contactless has a tight tap time and limited power, and PQC keys and signatures are large. Will the paper favor going online, fewer public-key operations, and lighter key-exchange-based card authentication (meaning further work on the C-8 specification)?
- Would EMVCo’s shareholders consider a contactless-first route? That means a quantum-safe C-8 2.0, promoted by the networks as the common kernel, with the contact flow later moving to the same model. Could the cost of migrating to PQC be the reason the networks finally converge on one contact and contactless kernel?
Starting this work now is cheap insurance. If a quantum computer never arrives, the industry still gets a modern, crypto-agile, converged kernel. If it does arrive, EMV is ready.
We look at how these questions could be answered in our follow-up whitepaper, Post-Quantum EMV: At a Crossroads.