DORA for fintech startups on AWS: a technical checklist

· 9 min read · compliance · fintech

DORA has applied since January 2025, and supervisors have moved from asking whether you are ready to asking for evidence. This is what the regulation expects from the engineering side of a small fintech on AWS, pillar by pillar.

Not legal advice. This post covers the engineering side of DORA. Whether you are in scope, which framework applies to you and how your supervisor reads it are questions for your compliance officer, legal counsel and competent authority.

DORA (Regulation (EU) 2022/2554) has applied since 17 January 2025 to banks, payment and e-money institutions, investment firms, crypto-asset service providers, insurers and most other EU financial entities. On AWS it translates into five areas of work: ICT risk management, incident reporting within 4 hours, 72 hours and one month, resilience testing, third-party risk with a register of information and exit plans, and information sharing.

Below: who is in scope, what each pillar means in engineering terms, and the AWS services and evidence that cover it. Every legal reference links to the source and was checked in November 2026.

Are you in scope?

If you hold an EU financial licence, almost certainly yes. If you are a SaaS or infrastructure company selling to financial entities, DORA reaches you through your customers' contracts, not directly.

Article 2 of DORA lists the financial entities covered: credit institutions, payment institutions, account information service providers, electronic money institutions, investment firms, crypto-asset service providers and issuers of asset-referenced tokens, crowdfunding service providers, fund managers, insurance and reinsurance undertakings, larger insurance intermediaries, trading venues and several market infrastructures. A few small categories are excluded, including insurance intermediaries that are micro, small or medium-sized enterprises.

You areHow DORA applies
Licensed fintech (payment institution, EMI, CASP, investment firm)Directly. You need the full framework, scaled to your size
Small payment or e-money institution exempted under PSD2 or the E-Money Directive, small non-interconnected investment firmDirectly, but under the simplified ICT risk management framework of Article 16
Microenterprise financial entity (fewer than 10 people and turnover or balance sheet up to EUR 2 million)Directly, with lighter obligations in several articles and no TLPT
SaaS, API or infrastructure vendor to financial entitiesIndirectly: Article 30 contract terms, audit and access rights, incident support, exit assistance
Designated critical ICT third-party provider (CTPP)Directly overseen by the European Supervisory Authorities

The ESAs published the first list of 19 critical ICT third-party providers on 18 November 2025. It includes Amazon Web Services EMEA SARL, Google Cloud and Microsoft. That puts AWS under direct EU oversight, but it does not move any of your obligations to AWS. Your register, your exit plan and your incident reports remain yours.

Pillar 1: ICT risk management

You need a documented ICT risk management framework, approved by your management body, that covers identification, protection, detection, response and recovery, backups and learning. Most of the technical substance is in Articles 6 to 14 and in Delegated Regulation (EU) 2024/1774, which also sets out the simplified framework.

The articles engineers should read are Article 9 (protection and prevention), Article 10 (detection), Article 11 (business continuity, tested at least yearly) and Article 12 (backups). Article 12 is specific: backups must be restored onto systems physically and logically segregated from the source, restoration must be tested periodically, and you must set recovery time and recovery point objectives.

DORA requirementAWS controlsEvidence to keep
Asset inventory and classification (Art. 8)AWS Config across all accounts, mandatory tags for owner and criticality, Resource ExplorerConfig aggregator export, list of systems supporting critical or important functions
Protection and access control (Art. 9)AWS Organizations with SCPs, IAM Identity Center with MFA, KMS encryption, no long-lived access keysSCP repository, quarterly access reviews, encryption status from Security Hub
Detection (Art. 10)Organisation CloudTrail to a log archive account, GuardDuty, Security Hub with CIS and Foundational Security Best Practices standardsAlert routing, on-call records, triage notes for findings
Continuity (Art. 11)Multi-AZ for every critical system, infrastructure as code, Resilience Hub assessmentsYearly continuity test reports, results of switchover tests
Backups (Art. 12)AWS Backup with cross-account and cross-region copies, vault lock in compliance mode, restore testingRestore test results with measured RTO and RPO
Change managementPull requests with review, CI/CD with OIDC, Config rules for driftMerge history, deployment logs

Our AWS disaster recovery guide covers the backup design that satisfies Article 12, including immutable copies in a separate account. Our SOC 2 on AWS checklist covers the identity, logging and detection baseline; most of it is reusable evidence for DORA.

Pillar 2: incident classification and reporting

For a major ICT-related incident, you send an initial notification within 4 hours of classifying it as major and no later than 24 hours after becoming aware of it, an intermediate report within 72 hours of the initial notification, and a final report within one month of the latest intermediate report.

The timelines come from Article 5 of Delegated Regulation (EU) 2025/301; the templates are in Implementing Regulation (EU) 2025/302. Whether an incident is major is decided with the criteria in Delegated Regulation (EU) 2024/1772: clients and counterparts affected, reputational impact, duration and service downtime, geographical spread, data losses, criticality of services affected and economic impact. If an incident affects your clients' financial interests, you must also tell them without undue delay (Article 19(3)).

ReportDeadlineWhat you need ready
Initial notification4 hours after classification as major, at most 24 hours after awarenessDetection time, classification decision, systems and clients affected
Intermediate report72 hours after the initial notificationImpact so far, actions taken, whether services are restored
Final reportOne month after the latest intermediate reportRoot cause, resolution, costs and losses

These are not NIS2's 24 and 72 hour deadlines, and several guides mix them up. DORA is the specific law for financial entities, so it applies in place of NIS2's reporting rules for them.

Four hours is tight for a small team at 3am. The engineering work that makes it possible: alerts that page a human, a classification checklist built from the 2024/1772 criteria, timestamps captured automatically (when GuardDuty or CloudWatch fired, when the incident channel opened), and a pre-filled template. Our incident response runbook template already has the roles and timeline log; add a DORA classification step and the regulator contact to it.

Pillar 3: digital operational resilience testing

Every financial entity other than a microenterprise needs a risk-based testing programme (microenterprises still test, on a lighter, risk-based basis), and systems supporting critical or important functions must be tested at least yearly. Threat-led penetration testing (TLPT) every three years applies only to entities their authority selects, and never to microenterprises.

Article 25 lists the tests: vulnerability assessments and scans, open source analysis, network security assessments, source code reviews where feasible, scenario-based, performance and end-to-end tests, and penetration testing. Article 26 sets TLPT at least every three years for entities identified by the competent authority based on their impact and systemic importance. A seed-stage payment institution is very unlikely to be selected.

Pillar 4: ICT third-party risk and the register of information

You must keep a register of every contract for ICT services, report it to your authority every year, include the mandatory contract terms of Article 30, and have tested exit strategies for providers supporting critical or important functions. For a fintech on AWS, AWS is the first line in that register and the hardest exit plan to write.

The register follows the 15 templates of Implementing Regulation (EU) 2024/2956. The cycle since 2026 uses a 31 December reference date, with national authorities forwarding registers to the ESAs by 31 March; each authority sets its own earlier deadline for entities. List every ICT provider, not only cloud: your KYC API, card processor, core banking SaaS, observability vendor and email provider.

On contracts, AWS offers a DORA Financial Services Addendum on request through your account manager, plus a DORA user guide and workbook in AWS Artifact. Request them early; they save weeks of legal back-and-forth.

The exit strategy (Article 28(8)) is where engineering matters. Regulators do not expect a startup to run hot in two clouds. They do expect a credible, tested plan. Ours usually contains:

If you are the vendor rather than the regulated entity, expect your financial customers to ask for audit rights, incident notification commitments, sub-processor lists and exit support in contracts. Having ISO 27001 or SOC 2 helps; our comparison of SOC 2 and ISO 27001 for EU startups explains which one financial buyers tend to accept.

Pillar 5: information sharing

Sharing cyber threat information with other financial entities is voluntary under Article 45. For a startup it usually means joining an existing sector arrangement and telling your authority you did.

The engineering side is small: GuardDuty threat intelligence and Security Hub integrations give you indicators; a shared arrangement gives you context. Voluntary notification of significant cyber threats to your authority is also possible under Article 19(2).

The simplified framework and microenterprises

Smaller entities get less, not nothing. The simplified framework of Article 16 replaces Articles 5 to 15 for exempted payment and e-money institutions, small non-interconnected investment firms and small pension funds, and microenterprises get proportional relief in several articles.

Under Article 16 and Title III of Delegated Regulation 2024/1774 you still need an information security policy, an inventory and classification of critical functions and ICT assets, a risk assessment, access control, logging, change management, backups and a tested business continuity plan. Microenterprises are defined in Article 3(60) as fewer than 10 people with annual turnover or balance sheet up to EUR 2 million; they are excluded from TLPT, and their yearly continuity tests do not have to include cyber-attack and switchover scenarios. Incident reporting and the register of information still apply in both cases.

In practice, the AWS baseline is the same either way. What changes is the volume of documentation and testing around it.

A 90-day order of work

Start with what a supervisor or an incident would expose first: incident reporting, backups and the register. Policy polish comes last.

  1. Weeks 1 to 2: list critical or important functions and the systems behind them; draft the register of information.
  2. Weeks 2 to 4: organisation CloudTrail, GuardDuty and Security Hub in every account; alerts to an on-call rota.
  3. Weeks 3 to 6: AWS Backup with cross-account copies and vault lock; first timed restore.
  4. Weeks 4 to 8: incident runbook with DORA classification and the 4h, 72h and one-month reports; a tabletop exercise.
  5. Weeks 6 to 10: exit strategy for AWS and other critical providers; request the AWS DORA addendum.
  6. Weeks 8 to 12: testing programme, first external pen test, management body approval of the framework.

The free uptime and SLA calculator helps turn your recovery objectives into numbers the management body can approve.

If you are a fintech getting ready for a supervisory review, our cloud audit checks your AWS estate against these pillars and gives you a prioritised gap list with the evidence you are missing. If you need the controls run every day, including restore tests, findings triage and on-call, our managed DevOps service does that work for you.

Frequently asked questions

When did DORA start to apply?

DORA, Regulation (EU) 2022/2554, has applied since 17 January 2025. Its technical standards followed in delegated and implementing regulations, including the incident reporting timelines in 2025/301 and the register of information templates in 2024/2956. Supervisors now expect evidence that the framework operates, not just that policies exist.

What are the DORA incident reporting deadlines?

For a major ICT-related incident: an initial notification within 4 hours of classifying it as major and no later than 24 hours after becoming aware of it, an intermediate report within 72 hours of the initial notification, and a final report within one month of the latest intermediate report. These come from Delegated Regulation (EU) 2025/301.

Does using AWS make a fintech DORA compliant?

No. AWS was designated a critical ICT third-party provider in November 2025 and is overseen by the ESAs, but your obligations stay with you: the register of information, contract terms, exit strategy, incident reporting and testing. AWS provides a DORA addendum and user guide that help with the contract and shared responsibility.

Do microenterprises have to comply with DORA?

Yes, with proportional relief. A microenterprise has fewer than 10 people and turnover or balance sheet up to EUR 2 million. It is excluded from threat-led penetration testing and has lighter testing and governance requirements, but incident reporting, backups and the register of information still apply.

Tell us what you need