Opsio - Cloud and AI Solutions
GDPREnforcement7 min read· 1,456 words

GDPR Fines Case Law: Patterns from €1bn+ Enforcement Actions and What CISOs Should Learn

Published: ·Updated: ·Reviewed by Opsio Engineering Team
Fredrik Karlsson

Group COO & CISO

Operational excellence, governance, and information security. Aligns technology, risk, and business outcomes in complex IT environments

GDPR Fines Case Law: Patterns from €1bn+ Enforcement Actions and What CISOs Should Learn

Cumulative GDPR fines crossed €5.6bn by the end of 2025 according to the EDPB's annual report and CMS's Enforcement Tracker — up from €1.6bn at the end of 2022. The headline numbers (Meta €1.2bn in May 2023, Amazon €746m in July 2021, Meta €405m on Instagram in September 2022, TikTok €345m in September 2023) get the press attention, but the patterns underneath them are what actually matter to CISOs and DPOs designing controls. Read enough decisions and the same defects keep recurring; the fines are large because of the controllers' scale, not because the underlying violations are exotic.

This piece surveys the ten largest GDPR enforcement actions to date, extracts the patterns, and translates them into design rules for security and privacy programmes. It is for CISOs and DPOs who already know the regulation and want to know which control failures get the biggest cheques attached.

The Ten Largest Fines as of Early 2026

YearControllerFineDPACore violation
2023Meta Platforms (Ireland)€1.2bnIrish DPC (binding EDPB decision)Art. 46(1) — US transfers without adequate safeguards post-Schrems II
2021Amazon Europe Core€746mLuxembourg CNPDBehavioural advertising without valid consent
2022Instagram (Meta)€405mIrish DPCArt. 8 children's data, public-by-default settings on minor accounts
2023TikTok€345mIrish DPCChildren's data, public-by-default, dark patterns
2023Meta (Facebook + Instagram)€390m combinedIrish DPCArt. 6(1)(b) contract-as-basis for behavioural ads
2021WhatsApp (Meta)€225mIrish DPCArt. 13/14 transparency failures
2024Uber€290mDutch APArt. 44 transfers of driver data to US without safeguards
2024LinkedIn (Microsoft)€310mIrish DPCBehavioural ads, basis selection, transparency
2022Google (Ireland)€90mCNILCookie consent dark patterns ("Reject all" not equally visible)
2020H&M€35mHamburg HmbBfDIEmployee surveillance — extensive recording of personal life)

Five patterns explain almost every entry on this list.

Pattern 1: Wrong Lawful Basis (Articles 6 and 9)

Three of the top ten fines (Meta 2023 €390m, Amazon 2021 €746m, LinkedIn 2024 €310m) turn on lawful-basis selection for behavioural advertising. The pattern is identical: controller treats personalised ads as part of the service contract under Article 6(1)(b), the EDPB or DPA finds that ads are not necessary for performance of the contract, and consent is required instead. The Bundeskartellamt v Meta CJEU ruling (Case C-252/21, July 2023) crystallised this in EU law — necessity for contract is interpreted strictly, and "freely given consent" requires a real alternative.

Design rule: any processing where the controller frames "the service" broadly to absorb the processing should be re-examined. If the user could be served the core product without the processing, contract is the wrong basis. The LIA path is not a free pass — fraud prevention and own-customer marketing are defensible legitimate interests under Recitals 47 and 49; cross-context behavioural advertising is generally not, after EDPB Guidelines 1/2024.

Free Expert Consultation

Need expert help with gdpr fines case law?

Our cloud architects can help you with gdpr fines case law — from strategy to implementation. Book a free 30-minute advisory call with no obligation.

Solution ArchitectAI ExpertSecurity SpecialistDevOps Engineer
50+ certified engineersAWS Advanced Partner24/7 support
Completely free — no obligationResponse within 24h

Pattern 2: International Transfers and Schrems II Failure

Meta €1.2bn (2023) and Uber €290m (2024) are both Article 44 transfer failures. Meta continued to transfer EEA user data to the US after Privacy Shield was invalidated by Schrems II in July 2020, without supplementary measures sufficient to address US surveillance access. Uber transferred driver data without an adequacy decision or SCCs supporting the specific transfer. Both cases pre-date the EU-US DPF adequacy decision of July 2023, but the design lesson stands: transfer mechanism gaps compound monthly. Each month of unlawful transfer is itself a violation, and quantum scales with the duration.

Design rule: transfers map and TIA must be living documents, not one-off compliance artefacts. When a transfer mechanism is invalidated (as Privacy Shield was), the controller has weeks, not years, to remediate. The DPF is currently valid but Schrems III is pending; controllers running US sub-processors should have SCC fallback ready and supplementary measures (encryption with keys held in EU) in place.

Pattern 3: Children's Data and Default Settings (Article 8 and Article 25)

Instagram €405m (2022) and TikTok €345m (2023) are both children's-data cases. The same defect appears in both: minors' accounts were public by default, exposing follower lists, contact details, and other data without active opt-in. Article 8 sets the children's-data threshold (16 unless member states lower it to 13); Article 25 requires data protection by design and by default, with default settings applying the most privacy-friendly options. Public-by-default for minor accounts is the textbook violation of Article 25(2).

Design rule: any product surface that handles minors' data must enumerate every default and verify the most privacy-protective setting is the default. Age gating must be robust enough to actually filter — self-declared age fields without verification have been criticised in DPA decisions. Where the user base may include minors, design as if it does.

Pattern 4: Transparency Failures (Articles 12, 13, 14)

WhatsApp €225m (2021) is the canonical Articles 13/14 case — privacy notices that did not adequately explain processing for service improvement, security, and integration with other Meta products. Several smaller fines across CNIL and AEPD enforcement follow the same pattern: the privacy notice exists, but is generic, layered behind too many clicks, or fails to identify the legitimate interest specifically.

Design rule: privacy notices must be processing-specific and accessible from the surface where the data is collected. Layered notices are acceptable per the EDPB Transparency Guidelines (WP260 rev.01) only if the top layer contains the essential information per Article 13(1)/(2). Generic "we may use your data for various purposes" language fails. Each processing activity in the RoPA should map to a transparency entry.

Pattern 5: Dark Patterns and Cookie Consent

Google €150m (2022), Facebook €60m (2022), Yahoo €10m (2024), Microsoft €60m (2022) are all CNIL cookie-consent fines under ePrivacy + GDPR. The fact pattern is consistent: "Accept all" is one click and prominent; "Reject all" is hidden behind multiple clicks or absent entirely. The CNIL position, now adopted by the EDPB Cookie Banner Taskforce report (January 2024), is unambiguous — refusal must be as easy as acceptance.

Design rule: every consent surface needs an accept/reject parity audit. Pre-ticked boxes, color-contrast bias toward "accept", multi-step refusal flows, and "legitimate interest" toggles for processing that requires consent are all enforcement triggers. The bar is now empirically established by 200+ DPA decisions across the EEA.

Pattern 6: Security and Article 32 / Article 33 Failures

Beyond the top ten, a long tail of breach-driven fines (British Airways £20m, Marriott £18.4m, Equifax-via-CNIL, several hospital ransomware cases) hinge on Article 32 security inadequacy and Article 33 breach notification timing. Common defects: missing MFA on privileged access, unpatched perimeter assets, weak encryption-at-rest, no segmentation between dev and prod, breach detected by third party rather than by the controller, 72-hour notification missed. The pattern is unspectacular and avoidable.

Design rule: Article 32 is not a checklist; it is a risk-aligned control set. The same controls that earn ISO 27001 certification cover the great majority of Article 32 expectations. Pair the Article 33 breach playbook with the SOC tooling — the 72-hour clock starts at awareness, not at investigation closure. A managed detection and response capability that surfaces incidents within hours is the difference between meeting and missing the deadline.

What the Patterns Mean for CISO and DPO Programmes

Six concrete control investments cover most of the historical fine surface: a well-documented LIA library covering every legitimate-interest claim; a transfer-mechanism inventory tied to the RoPA with quarterly re-evaluation; default-setting audits on all customer surfaces, particularly any where minors may appear; processing-specific transparency notices linked from the RoPA; a CMP and consent-surface audit against the EDPB Cookie Banner Taskforce checklist; and an Article 32/33 control set anchored in ISO 27001 with a 72-hour breach playbook integrated with the SOC stack.

None of these are exotic. The €5.6bn collected so far has not been collected for novel violations; it has been collected for the same six control gaps repeated across thousands of organisations.

How Opsio Helps

Opsio runs GDPR control assessments and remediation programmes for European enterprises, drawing on enforcement-trend analysis to focus investment on the controls actually generating fines. Our GDPR compliance services with Opsio cover lawful-basis review, transfer-mechanism mapping, default-settings audits, transparency-notice rewrites, CMP-implementation verification, and Article 32/33 readiness. We pair this with cloud security consulting and managed detection and response services so the breach-handling side of GDPR is staffed and instrumented, not just policy-documented.

About the Author

Fredrik Karlsson
Fredrik Karlsson

Group COO & CISO at Opsio

Operational excellence, governance, and information security. Aligns technology, risk, and business outcomes in complex IT environments

Editorial standards: This article was written by a certified practitioner and peer-reviewed by our engineering team. We update content quarterly to ensure technical accuracy. Opsio maintains editorial independence — we recommend solutions based on technical merit, not commercial relationships.