Although the main requirements will not come into force until December 2027, manufacturers must start preparing as early as 11 September. (Photo: Shutterstock)

Although the main requirements will not come into force until December 2027, manufacturers must start preparing as early as 11 September. (Photo: Shutterstock)

The main requirements of the Cyber Resilience Act will not come into force until December 2027. However, manufacturers of digital products will be required to report certain vulnerabilities and incidents from 11 September 2026. Who is affected, what needs to be reported, and how can organisations prepare? Answers based on the European Commission’s guidelines.

Does the Cyber Resilience Act (CRA) apply only to technology companies?

No. The regulation applies to companies that place products containing digital elements on the European market. A manufacturing company that sells a connected machine may, therefore, be affected, even if it does not present itself as a technology company.

Conversely, a bank, a trust company or a business that simply uses software purchased from suppliers is not normally regarded as a manufacturer under the CRA. However, these organisations may be subject to other legislation, such as NIS2, Dora or the GDPR.

Which products fall within the scope of the regulation?

The CRA may cover software, installable applications, components, operating systems, connected devices, industrial machinery and other hardware or software products capable of communicating with a device or a network.

The analysis must be carried out on a product-by-product basis. It is not enough simply to look at the sector or the size of the company.

Who is considered to be a manufacturer?

A manufacturer is generally a company that develops a product, or commissions its development, and markets it under its own name or brand.

A company may also assume the manufacturer’s responsibilities where it makes a substantial change to an existing product before placing it back on the market. Importers and distributors have their own obligations regarding verification and cooperation, but the primary obligation to notify, as set out in Article 14, rests with the manufacturer.

Are SaaS services affected?

Not automatically. The Commission is examining whether the remote processing forms part of the product.

This is the case where its absence would prevent the product from performing one of its functions and where the remote software has been developed by the manufacturer or under its supervision. This function may include authentication, synchronisation, configuration, control of a device or the automatic distribution of updates.

Remote processing carried out solely for the purpose of generating statistics or preparing for future developments is not normally part of the product. It may, however, need to be taken into account in the risk analysis.

Does using a cloud provider transfer liability?

No. When a manufacturer deploys its own software in an IaaS or PaaS environment, that software may still be considered a component of the product, even if the infrastructure is owned by a cloud service provider.

A standard SaaS application provided by a third party is not generally regarded as having been developed under the manufacturer’s responsibility. The manufacturer must, however, assess the risks associated with its integration and verify the supplier’s security assurances. In particular, the Commission recommends that service contracts include guarantees relating to vulnerability management and customer notification.

What will need to be declared from 11 September 2026?

The manufacturer must report two categories of events: actively exploited vulnerabilities in its product and serious incidents that have compromised the product’s security.

The discovery of a vulnerability does not, therefore, automatically trigger a report. There must be evidence to show that it is being actively exploited. It must, however, be dealt with as part of normal vulnerability management.

What happens when the vulnerability originates from a third party or open-source component?

If a vulnerability originating from a third-party library or component is actively being exploited in the final product, the manufacturer must report it.

The situation is different where the vulnerable code is not accessible within the product’s configuration or where there is no indication that it has been exploited within the product. In such cases, the mandatory reporting requirement is not triggered for that manufacturer. They must, however, address the risk and inform the manufacturer or maintainer of the component.

Organisations designated as “open-source software stewards” also have a specific obligation to report actively exploited vulnerabilities of which they are aware when they are involved in the development of the software in question.

When does the time limit start?

The time limit begins when the manufacturer, following an initial assessment, has a reasonable degree of certainty that a vulnerability is being actively exploited or that a serious incident has compromised its product.

An unconfirmed signal therefore does not necessarily trigger the timer. However, the company must assess it immediately. It cannot artificially prolong its analysis in order to delay reporting it.

The report may come from a customer, a security researcher, a regulatory body, the media, an employee or an internal monitoring system.

What deadlines will need to be met?

An initial alert containing limited information must be forwarded without undue delay and, at the latest, within 24 hours of becoming aware of the matter.

An enhanced notification must be provided within 72 hours. In the case of an actively exploited vulnerability, the final report must be submitted within 14 days of a corrective or mitigating measure becoming available. In the case of a serious incident, it must be submitted within one month of the 72-hour notification.

The company must, therefore, not have completed its investigation within 24 hours. The system provides for information to be passed on gradually as the investigation progresses.

Where should the declaration be submitted?

The notification is to be sent simultaneously to the CSIRT acting as the coordinator for the Member State in which the manufacturer’s principal place of business is situated, and to the European Union Agency for Cybersecurity (Enisa).

In practice, manufacturers will use the CRA’s single European reporting platform, which is due to become operational on 11 September 2026. National clarification is still awaited regarding the identity and precise role of the Luxembourg-based stakeholders within this system.

Does this apply to products already on the market?

Yes. The reporting requirement will apply from 11 September 2026 to all products falling within the scope of the CRA, including those placed on the market before the regulation comes into general effect in December 2027.

The Commission even specifies that the reporting obligation continues after the end of the product’s support period. A company will therefore need to maintain the capacity to receive, assess and report on certain older products that it no longer supports.

Will we need to register farms that were already known before 11 September?

No. The CRA does not require a retrospective disclosure where a manufacturer was already aware, prior to 11 September 2026, that a vulnerability was being actively exploited.

However, if the vulnerability was known before that date but its active exploitation was only discovered afterwards, it must be reported.

Will users need to be informed?

Yes. The manufacturer must inform the affected users and, where appropriate, all users.

This obligation does not necessarily require all technical details to be made public immediately. In the case of a product used in a sensitive environment, the information may be disclosed only to the customers concerned where its publication could facilitate further attacks. A wider disclosure may take place once the vulnerability has been rectified.

How can a business prepare?

It should start by drawing up an inventory of the products it sells, including legacy products, open-source components, cloud services and functions provided remotely.

It must then determine its legal role in relation to each product, identify the individuals authorised to classify an event, prepare notification templates and document the point at which it reaches a reasonable degree of certainty.

Contracts with developers, component suppliers and cloud service providers must ensure that information is reported back promptly enough. Finally, an internal procedure should be put in place to coordinate any parallel reporting requirements imposed by the CRA, NIS2, Dora or the GDPR.

The most useful test involves simulating the discovery of an actively exploited vulnerability on a Friday evening. The company must be able to quickly determine whether its product is affected, who makes the decision and who issues the statement within 24 hours.