Post-Quantum Cryptography for Banks DORA, NIS2, BSI

Quantum computing has been on the horizon for years. As the technology matures, however, that horizon is drawing ever closer, and it is becoming increasingly difficult to treat quantum computing as a distant technology topic.

Quantum computing is appearing more and more in technology strategies, cybersecurity discussions and the European regulatory landscape. This raises a practical question for banks: when does post-quantum cryptography move from being a topic to be monitored to one that requires action?

We work with many players from the banking and insurance sector. Given our expertise in regulatory requirements such as DORA and NIS2, we have therefore examined post-quantum cryptography from both a regulatory and a technical perspective.

Our conclusion: the regulatory clock is ticking and technological pressure is rising.

A large part of today's digital infrastructure is based on public-key cryptography. In simple terms, public-key cryptography uses a pair of mathematically linked keys, one public and one private, to exchange information securely and verify identities without first having to exchange a secret key. It protects confidential communication, authenticates systems and users, enables digital signatures and forms the foundation of trust for networked systems.

The problem is that sufficiently powerful quantum computers will one day break important public-key cryptographic algorithms that are in use today. The European Commission therefore describes the transition to post-quantum cryptography, or PQC for short, as necessary to protect Europe's digital infrastructure against future quantum attacks.

It would be easy to conclude that this only becomes a problem once such a quantum computer exists. However, that overlooks a significant danger.

Information that is encrypted today can be, and already is being, collected and stored with the aim of decrypting it as soon as the necessary quantum technologies become available. The European Commission explicitly warns that data which must remain confidential in the long term may already be at risk today, and points out that so-called "harvest now, decrypt later" attacks are probably already taking place.

For banks, this changes the question. Instead of asking "When will quantum computers be powerful enough?", banks should increasingly ask:

"Which of our data must remain confidential in the future?", "Which systems protect this data?" and "Where should we already be preparing the transition to quantum-safe methods today?"

To make matters worse, cryptography cannot simply be replaced overnight. It is embedded in applications, infrastructure, certificates, interfaces and third-party services. Preparing for PQC is therefore as much a migration challenge as a cryptographic one.

Looking at PQC from a regulatory perspective, one thing becomes clear: there is no single European regulation that tells banks precisely which post-quantum algorithm they must implement and when.

Instead, several regulatory and policy instruments are increasingly interlocking.

For banks, DORA provides the sector-specific framework for ICT risk. NIS2 creates the broader European context for cybersecurity and governance; where DORA regulates corresponding requirements on a sector-specific basis, these DORA requirements apply to the financial entities concerned in place of the corresponding NIS2 provisions.

The EU's PQC Recommendation and the joint guideline from BMDS, BMI and BSI on the national implementation of the EU roadmap, published on 12 October 2026, make the direction of the transition more concrete. Recommendations and guidelines must be distinguished from directly applicable legal obligations.

The following milestones for organisations on the path to Quantum-Readiness can be derived in particular from the BSI roadmap:

Whether these tight timelines are truly necessary depends on the pace at which the technology is developing. Quantum computing is a double-edged sword which, alongside the risks, also holds considerable potential. As a result, a wide range of companies and institutions are driving technological development. We saw the emergence of an economic ecosystem around quantum computing, for example, at Bitkom's AI, Data and Quantum (AIDAQ) conference in Berlin. Technological development is expected to accelerate further in the coming years, driven by AI and by funding such as that announced in the EU Quantum Act. Rapid technological development, however, also means growing technological risk and increasing pressure to act.

Post-quantum cryptography (PQC) is necessary regardless of regulation: the technological risk exists today. Attackers can steal and store sensitive data now in order to decrypt it later using powerful quantum computers.

One possible scenario illustrates the risk: if customer, payment or transaction data is stolen today and decrypted in 2031, for example, the disclosure of confidential information cannot be prevented retrospectively. Attacks on digital identities and signatures are a separate matter: these may become possible if the cryptographic methods used for them are broken by powerful quantum computers. Both attack paths, data decryption and signature forgery, can shake trust in digital processes.

Possible worst-case scenario for a bank: attackers decrypt intercepted account balances and payment histories, exposing the bank to liability claims and customer attrition as a result of the loss of trust. At the same time, the attackers break quantum-vulnerable certificate and signature methods: using forged identities, they initiate fraudulent transfers. The bank suffers financial losses and has to block access and interrupt payment processes. High recovery costs and lasting reputational damage aggravate the consequences.

Migration is therefore not a pure compliance task, but a technical and economic necessity.

To avoid such nightmare scenarios, it is important that banks take action today in order to be prepared. This does not mean replacing every encryption method immediately. The first step is to build an overview of potential vulnerabilities and the measures required.

The aim of the preparation phase is to understand the existing cryptographic landscape, identify the most significant threats and build the capability to migrate in a controlled manner. The starting point is, somewhat surprisingly, something familiar: a cryptographic inventory.

Such an inventory is not created by a technical stocktake alone: it must first be clear who is responsible for the collection and who coordinates the further measures. The BSI guideline offers orientation here; from it, a possible approach for banks to prepare for PQC can be derived. The cryptographic inventory forms the basis for a targeted transition to quantum-safe methods.

A record with various pieces of information should be created for each component in the IT landscape:

An entry in the cryptographic inventory could look like this, for example:

The more complete the cryptographic inventory, the lower the risk that potential vulnerabilities will be overlooked during the transition. All IT components should therefore be taken into account.

Under DORA, a bank remains responsible for meeting its regulatory obligations even when using external ICT services. Components and services obtained from third-party providers should therefore also be included in the cryptographic stocktake.

Creating the cryptographic inventory can be broken down into 4 steps:

PQC preparation first means understanding the cryptographic landscape, assessing risks and being able to plan and steer the migration. The longer-term target picture goes further, covering implementation and permanent embedding in processes. This includes the following:

• all critical cryptographic methods are inventoried
• systems have been classified by quantum risk
• high-risk systems have been migrated to PQC or hybrid methods
• procurement and architecture will in future permit only PQC-capable solutions
• suppliers demonstrably have PQC roadmaps
• PQC is a firm component of architecture, risk and security processes