GDPR fines: what reduces them and what raises them

The European Data Protection Board has published draft guidelines on how supervisory authorities calculate GDPR fines. For managing directors and IT leads, the most useful part is how a company's own conduct after a breach changes the result.

Sandra Leistner4 min read

Illustration: A scale where coins are outweighed by a fire extinguisher and a measuring tape, with a ruler in front.

What the new EDPB guidelines on fine calculation cover

The European Data Protection Board (EDPB) published Guidelines 04/2026 on the calculation of fines in September 2026, as reported by Dr. Datenschutz. A public consultation runs until 13 November 2026, so details may still change before the final version. The underlying question will stay relevant: how does an authority arrive at a number, and what can a company influence?

The method has several steps. The authority first decides which processing operations and which infringements are sanctioned separately. It then sets a starting amount based on the gravity of the infringement and the company's turnover. Aggravating and mitigating circumstances follow, and finally the statutory maximum and the proportionality of the result are checked.

The guidelines contain no fixed price list for particular breaches. They also leave untouched the discretion that Art. 83 GDPR gives to each authority, and national procedural law still governs how a fine proceeding runs. The calculation becomes more uniform and transparent, while the outcome still depends on the case.

Which factors decide the amount of a fine

Gravity carries the most weight. It depends on the type of data affected, the number of people concerned, the duration of the breach and the degree of fault. A single configuration error without visible harm is judged differently from years of processing without a valid legal basis. Turnover serves as a reference point so that the deterrent effect is comparable for companies of different sizes.

The second level is the company's behaviour before, during and after the incident. According to the source, these points commonly mitigate a fine:

  • Timely, consistent and reliable cooperation with the authority
  • Reporting a breach under Art. 33 GDPR on the company's own initiative
  • Informing affected people under Art. 34 GDPR and blocking compromised access
  • Documented technical and organisational corrections and deletion of unlawfully collected data
  • Records of processing, data protection impact assessments and training records
  • Certifications and codes of conduct under Art. 40 and Art. 42 GDPR that cover the process concerned

Aggravating factors include earlier infringements or warnings, intent, financial gain from the breach, special categories of data under Art. 9 and Art. 10 GDPR, late or contradictory statements, and continuing a processing after an order to stop. Mitigating circumstances must be proven, and a bare claim does not count. Gaps in documentation can mean that safeguards which really existed are ignored.

How predictable are GDPR fines now?

Predictability improves in part. A common method makes decisions easier to follow and to challenge, and a company that knows the steps can argue about a specific one, such as the gravity rating. Realistically it can forecast a range but not an exact sum.

Several limits still apply. Authorities in the member states set different priorities and work with different resources and procedures. Courts keep refining the framework: the Court of Justice in Deutsche Wohnen held that a fine requires fault, yet a company can be held liable directly without a specific person being identified.

The 1&1 Telecom case shows how far review can go. The Regional Court of Bonn confirmed the infringement but found the authority's calculation inappropriate and reduced the fine from 9.55 million euros to 900,000 euros, taking the company's cooperation into account. For corporate groups, the upper limit follows the turnover of the whole economic unit.

What to document after a data protection incident

An incident rarely becomes expensive simply because it happened. The cost grows when the response is unstructured. A complete timeline helps: when the incident was detected, who was informed, which risk assessment informed the decision on reporting, and why that decision was taken. Art. 33(5) GDPR also requires a reasoned record where a company decides not to report.

  1. Record each corrective measure with a date and a named owner.
  2. Check whether the same flaw exists elsewhere and adjust processes and permission concepts.
  3. Name one person responsible for contact with the authority.
  4. Agree statements internally before they are sent.
  5. Write these steps into a response plan so they can run without delay.

Repeated infringements weigh heavily, so evidence that the root cause was fixed pays off later. The same logic applies to communication, because contradictory answers count against a company.

Using AI with company documents under these rules

souveraen.ai is an AI platform for company knowledge from clouditiv GmbH, based in Auerbach, Germany. It cites the document, version and page for each answer, and it reads connected sources such as SharePoint or file servers read-only with the existing permissions.

Agents prepare tasks, and anything that writes to a system waits for human approval. Administration offers roles, workspaces and a usage overview. The platform runs On Demand, hosted in Germany, or on-premise as Private Spark in the company's own network. Holders of professional secrets can sign a confidentiality undertaking under section 203 of the German Criminal Code. More on the topic is on the page about GDPR-compliant AI.

Sources

More articles

All articles

How safe are AI API routes via resellers and routers?

Security firm Dragos sent about two million requests with fake credentials to AI providers and found compromised services and resellers reading customer prompts. Companies that use AI through third parties need to know who sits on the path of their data.

Data protection4 min

Run Text Generation Inference with Docker on Your Server

Text Generation Inference (TGI) serves open language models from a single Docker container on your own hardware. This tutorial starts it, sends requests over HTTP and keeps prompts inside your network.

Tutorial4 min

Cloud outage liability: who answers when Microsoft fails

On 12 November 2026 a software fault took Exchange Online and Outlook offline for about 36 hours. A legal analysis published shortly afterwards argues that responsibility for such outages often lies with the operator rather than the provider.

Regulation4 min

Test it with your own documents.

Free for 3 days, then from €24.90 per user per month. Cancellable monthly.

Free for 3 days, then from €24.90 per user per month.Free for 3 days
Try free for 3 days