Skip to content

What the EU Cyber Resilience Act Requires from Software Development Teams

EU CRA

Published on

The EU Cyber Resilience Act (CRA) requires software teams to ship products with no known exploitable vulnerabilities, fix flaws for at least five years and report exploited vulnerabilities within 24 hours. It applies to any company that sells software or connected hardware in the EU, wherever that company is based.

24 hoursto report an exploited vulnerability. Live since September 11, 2026
Dec 11, 2027every other requirement applies
5 yearsminimum support period for security fixes
€15M or 2.5%top fine, whichever is higher, of worldwide annual turnover
  • Developers carry most of the work: secure design, secure defaults, a software bill of materials (SBOM) and regular security testing.
  • The CRA does not name developer training. It sets outcomes that only trained developers can deliver.

What is the EU Cyber Resilience Act?

The Cyber Resilience Act is an EU law that makes cybersecurity a condition of selling digital products in Europe. Its formal name is Regulation (EU) 2024/2847 and it entered into force on December 10, 2024.

The CRA covers "products with digital elements". That means any software or hardware with a direct or indirect data connection. A compliant product carries the CE mark, the same mark used for product safety.

The law puts the duty on the manufacturer. If your company builds the software and sells it under its own name, your company is the manufacturer.

Which software teams are in scope?

Your team is in scope if it builds software that customers install or that ships inside a device sold in the EU. Company location does not matter. A US firm selling to one EU customer must comply.

In scope

Desktop and mobile apps, operating systems and firmware

Connected hardware such as routers, smart home devices and industrial controllers

Cloud back ends that a product needs to perform its functions

Open-source software supplied as part of a commercial activity

Out of scope

Web apps used only through a browser, according to the Commission's July 27, 2026 guidance

Medical devices, motor vehicles and civil aviation products, which have their own rules

Open-source software shared with no commercial activity, according to the European Commission

What the CRA covers and what it leaves to other rules.

Most products can self-certify. The Commission's conformity assessment page reserves third-party assessment for higher-risk categories such as operating systems, firewalls and smartcards.

What must developers build into the product?

Developers must build 13 security properties into every product, based on a documented risk assessment. Annex I of the CRA lists them. They describe outcomes, not tools, so each one becomes a design and coding decision.

CRA requirement (Annex I, Part I) What it means for a development team
No known exploitable vulnerabilities at release Scan code and dependencies before every release and fix what you find
Secure by default configuration No default passwords, no open debug ports, least privilege out of the box
Protection from unauthorized access Sound authentication, authorization and session handling
Confidentiality and integrity of data Current encryption in transit and at rest, plus input validation and integrity checks
Data minimization Collect and process only the data the product needs
Availability under attack Rate limits, resource controls and resilience to denial of service
Limited attack surface Remove unused interfaces, services and dependencies
Security logging Record security events, with a user opt-out
Secure updates A signed, automatic update channel that users can postpone

These are the same weaknesses that fill the OWASP Top 10: Broken Access Control, Insecure Defaults and Vulnerable Components. The CRA turns them from best practice into a legal test.

What must teams do about vulnerabilities after release?

Teams must find, fix and disclose vulnerabilities for the whole support period, which is at least five years unless the product's expected use is shorter. Annex I, Part II sets eight duties. The main ones for developers are:

  1. Keep a machine-readable SBOM that covers at least top-level dependencies.
  2. Fix vulnerabilities without delay and ship security updates free of charge.
  3. Test and review product security regularly.
  4. Publish a coordinated vulnerability disclosure policy with a contact address.
  5. Disclose fixed vulnerabilities with enough detail for users to act.

Third-party code counts. Article 13 requires due diligence on every component you integrate, including open source.

Reporting is the part already in force. Since September 11, 2026, Article 14 requires manufacturers to report an actively exploited vulnerability or a severe incident on this clock:

Article 14 reporting clock
24 hours
 
Early warning that the event occurred
72 hours
 
Full notification with an initial assessment and mitigation
14 days
 
Final report for a vulnerability, counted from the day a fix is available
1 month
 
Final report for a severe incident, counted from the 72-hour notification

Bar lengths are illustrative, not to scale.

Reports go to the national incident response team and the EU cybersecurity agency ENISA through a single reporting platform. The duty covers products already on the market, not only new releases.

When do the CRA deadlines apply?

Two dates matter: September 11, 2026 for reporting and December 11, 2027 for everything else. The European Commission's implementation page lists both, and no delay to either has been enacted.

DoneDecember 10, 2024CRA enters into force
Live nowSeptember 11, 2026Reporting of exploited vulnerabilities and severe incidents
NextDecember 11, 2027Secure design, vulnerability handling, SBOM, documentation and CE marking

The three dates that define the CRA.

Products placed on the EU market from December 11, 2027 must meet every requirement. Products already on sale before that date fall under the full rules only if they are substantially modified. The reporting duty applies to them regardless.

The harmonised standards that let manufacturers prove compliance are running late. The Commission moved the first standards deadlines to October 31 and December 31, 2026. Teams that wait for final standards will run out of time. The Annex I requirements are already fixed in law, so start there.

What are the penalties for non-compliance?

The top fine is €15 million or 2.5% of worldwide annual turnover, whichever is higher. Article 64 applies it to breaches of the Annex I requirements, the manufacturer obligations and the reporting duty.

Maximum fines, Article 64
Annex I requirements, manufacturer obligations and reporting€15M or 2.5%
 
Other obligations under the CRA€10M or 2%
 
Misleading information to authorities€5M or 1%
 

Each cap is the fixed amount or the share of worldwide annual turnover, whichever is higher.

National market surveillance authorities can also order a product withdrawn or recalled.

Does the CRA require developer security training?

No. The CRA does not mention developer training. It requires results that depend on developer skill.

A product with "no known exploitable vulnerabilities" and "secure by default configuration" comes from developers who know how to write and review secure code. The technical file must also show a risk assessment, test reports and a vulnerability handling process. Each needs people who understand the work.

Our view

Training is the cheapest control a manufacturer can evidence. A scanner finds a flaw after a developer writes it. A trained developer does not write it.

How do we train teams for the CRA?

We built a CRA learning path of 25 lessons and 230 minutes. It pairs the regulation with the secure coding skills each requirement depends on.

CRA learning path

6 lessons on the law (30 minutes)19 lessons on secure development practice

The Security Journey CRA learning path: 25 lessons, 230 minutes.

Six lessons explain the law itself: purpose, product classification, role determination, vulnerability handling, core processes and obligations. They run 30 minutes in total and stand alone as an Essentials path for compliance, product and legal teams.

The other 19 lessons teach the practice behind Annex I. Developers work through:

  • Secure by default configuration and security from the start
  • Broken access control and denial of service, with hands-on exercises
  • The secure development lifecycle and securing a release
  • Dependency management and a two-part SBOM lesson
  • Vulnerability scanning, continuous monitoring and supply chain integrity checks

We designed the path this way for one reason. Reading the regulation tells a developer what the EU expects. It does not teach them to fix an access control flaw.

Frequently asked questions

Does the CRA apply to US software companies?

Yes, if they sell products with digital elements in the EU. The law follows the product, not the company's location.

Does the CRA apply to SaaS?

A web app used only through a browser is generally out of scope. A cloud service is in scope when an installed product or device needs it to function.

Is an SBOM mandatory under the CRA?

Yes. Manufacturers must keep a machine-readable SBOM covering at least top-level dependencies. They do not have to publish it, but authorities can request it.

How long must we provide security updates?

At least five years, unless the product's expected use is shorter. Each update must stay available for ten years or the rest of the support period, whichever is longer.

Does the CRA apply to open-source software?

Only when it is supplied in the course of a commercial activity. Foundations that steward commercial open-source projects have lighter duties and are exempt from fines.

Next step

See the CRA learning path

Ask us for a walkthrough of the CRA learning path and see how it maps to your product teams.

Request a walkthrough

This post is general information, not legal advice.