On 3 June, the EBA, ESMA and EIOPA published their first consolidated report on major ICT incidents under DORA. In 2025, 3,383 incidents were reported, averaging 0.18 per financial entity. Credit and payment sectors accounted for the majority. One third of disruptions had cross-border consequences. System failures and third-party provider events were the most common cause. Pure cyberattacks accounted for only 10 percent, but the ESAs stress that AI-driven threats require heightened preparedness.
Novudom Weekly – Week 24, 2026
This week's compliance landscape is shaped by three signals. The EU's financial supervisory authorities publish the first data-driven picture of ICT disruptions under DORA, pointing more to system failures and third-party risks than cyberattacks. The AI Act's transparency requirements became concrete with the final Code of Practice for labelling. And South Korea's record fine against Coupang confirms that inadequate basic controls cost more than sophisticated attacks.
First incident report reveals 3,383 major ICT disruptions in EU financial sector
The report shows that the greatest operational risk is not hackers but your own and your suppliers' system failures. Review contracts and SLAs with ICT providers. Ensure you have a tested incident process for third-party-related disruptions.
EU publishes final Code of Practice for labelling AI-generated content
The European Commission published the final Code of Practice for labelling AI-generated content on 10 June. The code provides guidance on how providers and deployers of generative AI should meet the transparency requirements in AI Act Article 50, applicable from 2 August 2026. The code introduces EU icons for AI labelling, promotes open standards, and aims to reduce compliance costs. Deepfakes and AI-generated text on matters of public interest must be clearly labelled.
Map where in your organisation AI-generated content is produced: marketing, customer communications, reports. Ensure labelling procedures are in place by 2 August. The cost of implementing labelling now is a fraction of a penalty.
Coupang fined USD 409M after breach affecting 34 million accounts
South Korea's data protection authority PIPC issued the country's largest-ever data protection fine on 11 June: KRW 624.7 billion (USD 409M) against e-commerce platform Coupang. A former employee accessed personal data from 34 million accounts over several months without detection. Of the fine, KRW 423.6 billion relates to leaked personal data and KRW 201.1 billion to unauthorised data collection. The authority's chair stated the breach was caused not by advanced hacking but by inadequate basic security management.
The fine confirms the trend: regulators hit harder on basic control failures than on sophisticated attacks. Privileged access and insider threats require monitoring. Verify that access management and logging of customer data access works in practice, not just on paper.
CISA launches CI Fortify: critical infrastructure must withstand isolation during state-sponsored cyberattack
The US cybersecurity agency CISA launched the CI Fortify initiative in May, urging critical infrastructure operators to plan for a scenario where their OT networks are compromised and telecom connections, internet, and vendor links are down. The guidance rests on two pillars: isolation (proactive disconnection from third-party networks) and recovery (documentation, backup, and transition to manual operations). CISA is already conducting pilot assessments.
CI Fortify is the US government's formal acknowledgement that destructive state-sponsored attacks on infrastructure are a realistic scenario. The same threat applies to Swedish organisations. Test whether your organisation can maintain critical functions without digital connectivity. This is the type of scenario planning the cybersecurity law requires.
Swedish government convenes industry and authorities for AI cybersecurity dialogue
On 4 June, the Minister for Civil Defence gathered representatives from government agencies, industry, academia, and the cybersecurity sector for a roundtable on how Sweden should address cybersecurity challenges arising from rapid AI development. The meeting is part of the government's updated action plan for the national cybersecurity strategy 2025 to 2029, now comprising 91 measures. On 8 to 9 June, Minister Erik Slottner participated in the EU Council meeting on telecoms and digitalisation and the D9+ cooperation in Luxembourg.
AI cybersecurity is now an explicit priority at government level. Organisations that can demonstrate a plan for AI-related risks position themselves more strongly, both in regulatory compliance and public procurement.
European Commission refers seven member states for failing to transpose NIS2
In May 2026, the European Commission referred seven member states to the Court of Justice of the EU for failing to transpose the NIS2 directive into national law within the set deadline. The first compliance audit deadline of 30 June is approaching, when essential entities in most countries must have completed their first formal compliance review. 21 of 27 member states have now transposed the directive. The NIS2 Cooperation Group adopted common incident reporting templates on 26 May, which the Commission plans to make binding through an implementing act.
Sweden transposed on time, but having the law in place is not enough. Has your organisation completed the first review against the cybersecurity law's requirements? 30 June is not an abstract EU deadline but an operational checkpoint.
Threat report: nation-state attacks on infrastructure double while ransomware declines
Waterfall Security's 2026 threat report shows that publicly reported cyber breaches with physical consequences in critical infrastructure fell 25 percent to 57 incidents. But behind the decline lies a shift: nation-state actors and hacktivists doubled their attacks, and the majority targeted critical infrastructure. Ransomware groups were affected by temporary factors, but the threat landscape is shifting from financial crime toward geopolitically motivated operations.
Declining ransomware statistics can create false comfort. Nation-state actors pursue different objectives than ransom: disruption, intelligence collection, and sabotage. Organisations in essential services should review their threat model rather than sizing their defences against yesterday's attacker.
Weekly reflection
This week shows that regulations are maturing from principles to measurable practice. DORA's incident report provides the first factual basis for benchmarking. The AI Act's labelling code turns transparency into a concrete implementation question. And the Coupang fine is a reminder that the costliest failures rarely involve advanced technology. The maturation of regulatory frameworks means that 'we are working on it' is no longer sufficient as an answer to the board. What is needed now is documented status, tested processes, and measurable results.