Whitepaper

Post-Quantum EMV: At a Crossroads

Our first whitepaper asked EMVCo five questions. This one looks at how they could be answered.

Where our first whitepaper left us

In our earlier article, Post-Quantum Readiness of EMV: Five Questions for EMVCo, we asked what EMVCo’s announced third paper would commit to. We know these are hard questions. This article looks at the possible directions, and at why planning should start before the answers are final.

The problem in short

The starting point is well known in the industry. and is not threatened by quantum computers. Offline card authentication relies on RSA and elliptic curves, which are. deprecates RSA-2048 after 2030 and disallows RSA and elliptic curves entirely after 2035. points to the same window, starting with the highest-risk uses, with finance .

Removing offline approval does not remove the problem. Open-loop transit still has offline data authentication at its core, even though the fare is authorised online afterwards. In much of Europe , including on transactions that are then approved online, and the enciphered PIN is protected with the card’s RSA key. In addition, the EMV Contactless Kernel (C-8) to authenticate the card and encrypt the exchange between card and terminal, regardless of whether the payment is then approved online or offline.

Upgrading the cryptography is no small step either. As we set out below, moving to post-quantum algorithms would change core security functions of EMV, multiply the amount of data exchanged between card and terminal, and require yet another large migration across the installed base.

Doing nothing is not a long-term option either. Within about four years, the RSA keys at the core of EMV fall into the category NIST’s draft plan deprecates, while . Without a migration path, issuers and acquirers have no credible modernisation plan to show their supervisors. And if that path cannot be found within EMV, the industry will look for one outside it.

That leaves the question of what the industry can do. We see two routes, each with its own challenges.

Route 1: remove offline

If every transaction were approved online and every PIN verified online, cards would no longer need public-key cryptography to prove they are genuine, and most of the quantum exposure in the card-to-terminal exchange would disappear without the need for new algorithms.

This route has gaps in four areas:

  • Transit. authenticate the card offline so the gate opens in a fraction of a second, and authorise later. Online-only operation means slower gates or more first-ride risk for operators.
  • Resilience. European policy is moving towards more offline capability, not less. Most recently, have publicly committed to offline card payments as a contingency measure.
  • PIN. Many markets rely on offline PIN, including when the transaction itself is authorised online by the issuer, so moving every transaction online would not remove it. Given the contingency arrangements above, offline PIN is better seen as a building block of resilience than as a legacy feature: a PIN cannot be checked online when the network is down.
  • C-8. Its privacy protection relies on elliptic-curve key agreement on every transaction and would still need a post-quantum replacement.

The open question for this route is which offline use cases must move to post-quantum cryptography, and which can remain on cryptography that regulators and the public will soon regard as outdated.

Route 2: bring post-quantum cryptography into the specifications

. What is missing is a version of the EMV specifications that uses them. The challenges are substantial:

  • Data volume. . The data a terminal reads to authenticate a card offline grows from under 1 kB today to about 10 kB with a direct substitution, and to around 4 kB with the most compact options.
  • Speed. A contactless tap allows about half a second. At the 106 kbit/s rate contactless EMV uses today, 10 kB takes about three-quarters of a second to transfer, before any computation.
  • Card computation. Post-quantum operations need more memory and dedicated hardware on the card, and some, such as ML-DSA signing, can take unpredictably long. That makes a direct one-for-one replacement of today’s algorithms unsuitable for a contactless tap, and calls for a different protocol design.
  • Readers. , up to 848 kbit/s, which would carry the same data in under a tenth of a second. Using them requires support in readers and cards, new firmware or new hardware, and a new full chain of approvals (L1, L2 and L3) across the installed base.
  • Crypto agility. The new specification should let algorithms and key sizes change through new keys, terminal software updates and the normal card reissue cycle, so that a future change does not need another migration on this scale.

. C-8 already authenticates the card through key agreement rather than a signature, and that pattern carries over to post-quantum key exchange. Compact signatures for certificates, and reading offline data only when a transaction will actually be approved offline, reduce the data at the tap.

The open questions for this route include how the data exchange will fit within the time a contactless tap allows, who decides on the bit rate, how old and new cards and terminals will work together during the transition, whether to combine classical and post-quantum algorithms for a period, which algorithm fits which role, and how laboratories and test tools will be ready in time. The list is unlikely to be complete. What is not open is interoperability. Today any card works at any terminal, and the industry will expect nothing less from a post-quantum EMV, so interoperability has to be designed in from the start.

Why the work should start now

Whichever route is chosen, adoption takes years. EMVCo itself estimates that terminal infrastructure needs at least ten years to implement a change, so work that starts now already runs past 2035.

The open questions fall into three groups:

  1. Regulatory. Which exemptions regulators would accept and on which grounds, for which use cases, and with which transition periods.
  2. Business. What the migration will cost, and how the cost is shared among schemes, issuers, acquirers and merchants, including where offline payments should remain and where they can be removed.
  3. Technology. What a post-quantum version of the EMV specifications must define: offline card authentication, PIN encipherment, C-8’s privacy protection, contactless bit rates, the approvals that follow, and how all this should relate to contact. Shouldn’t we finally use the same transaction flow for contact and contactless? This is where answers from EMVCo are needed first, because the cost calculations and the regulatory discussion depend on them.

All three discussions need to start now and in parallel. The regulatory and business questions are for regulators, schemes, issuers and acquirers to settle. The technical question is EMVCo’s: defining the specifications that enable secure and interoperable payments is its mission, and there is work to be done there.