Future: deadlines, forecasts, scenarios
This section tracks what can be expected in cybersecurity between 2027 and 2030. The calendar lists the binding dates of laws, standards and product support; the log records who said what about the future and how it can later be measured. They are followed by the questions experts disagree on, the main scenario studies, and the outlook for Hungary and Central and Eastern Europe, and finally a look back at the predictions made for 2026. Every item went through a verification gate: anything not found in the original source was left out, and inaccurate items were corrected.
Deadline calendar
Status: fixed — set by law, a published standard or a vendor announcement; pending — the date is expected but not yet set or adopted; suspended — the deadline has been postponed or halted.
2026
- 2 Dec 2026
fixed EU Artificial Intelligence Act (Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744): the new prohibitions under Article 5 (e.g. generating intimate images without consent) apply; the grace period for the labelling obligation under Article 50(2) expires for generative systems placed on the market before 2026-08-02
Affects: Providers and deployers of AI systems in the EU · source · verification source
- 9 Dec 2026
fixed New EU Product Liability Directive (Directive (EU) 2024/2853): applies to products placed on the market or put into service after 9 December 2026; software (including AI systems) counts as a product, except non-commercial free and open-source software
Affects: software and hardware manufacturers, importers · source
- 11 Dec 2026
fixed Cyber Resilience Act (CRA, Regulation (EU) 2024/2847), Article 35(2): by this date Member States should endeavour to ensure that a sufficient number of notified conformity assessment bodies are available in the Union
Affects: member states, conformity assessment bodies · source
- 24 Dec 2026
fixed eIDAS 2 (Regulation (EU) 2024/1183), Article 5a(1): every Member State must make at least one EUDI Wallet available (within 24 months of the entry into force of the implementing acts; the German app is planned for early 2027)
Affects: member states, citizens · source · verification source
- 31 Dec 2026
fixed Hungary: deadline for cybersecurity training of executives and information security officers appointed before 2025-12-31 (Ministry of Energy (EM) Decree 17/2025 (VII. 24.))
Affects: Executives and information security officers of organisations within the scope of the Hungarian Cybersecurity Act (Kibertv.) · source · verification source
- 31 Dec 2026
fixed EU PQC roadmap (NIS Cooperation Group, 2025-06-23): by the end of 2026 every Member State must begin the transition to post-quantum cryptography (non-binding recommendation; the deadline for critical infrastructure is the end of 2030)
Affects: member states, critical infrastructure operators · source · verification source
2027
- 2027 (during the year, no date)
pending Expected adoption of the directive amending NIS2 (COM(2026) 13), followed by a 12-month transposition period — NOT a fixed deadline; as of 2026-11-10 still at committee stage
Proposal (COM(2026) 13, 2026/0012(COD)): according to the summary provided, it was at parliamentary committee stage on 10 November 2026, no adoption date; we did not verify this on an official EU website.
Affects: Entities subject to NIS2, Member States · source · secondary source
- Autumn 2027 (date not set)
pending Polish parliamentary election (no later than November 2027); the Italian parliamentary election is due no later than December 2027; the Spanish election has been brought forward to 29 November 2026
The dates of the Polish and Italian elections have not been set (according to secondary sources, 11 November 2027 at the latest and December 2027 respectively); the Spanish election drops out of 2027, because on 5 October 2026 Sánchez called an early election for 29 November 2026 (the link confirms this).
Affects: electoral bodies, political parties, media · source · verification source · secondary source
- 1 Jan 2027
pending Hungary: planned transfer of the cybersecurity tasks and powers of the Supervisory Authority for Regulatory Affairs (SZTFH) to the minister responsible for cybersecurity (science and technology); registered organisations would not have to register again
Draft put out for public consultation by the Ministry of Science and Technology on 6 November 2026; the consultation ran until 14 November. There is as yet no promulgated legislation.
Affects: registered organisations and auditors under the cybersecurity supervision of the Supervisory Authority for Regulatory Affairs (SZTFH) (the number of organisations is unverified) · source · verification source
- 1 Jan 2027
fixed NSA CNSA 2.0: new acquisitions of national security systems must be CNSA 2.0-compliant
Affects: suppliers of US national security systems · source · verification source
- 12 Jan 2027
fixed End of support for Windows Server 2016
Affects: corporate and public administration server operators · source
- 3 Apr 2027
fixed Poland: deadline for meeting the obligations under the amended KSC Act (NIS2 transposition) for entities already covered when it enters into force (ISMS, incident handling, S46 connection)
Affects: Polish essential and important entities · source · verification source
- 18 Apr 2027
fixed French presidential election, first round (second round: 2 May 2027)
Affects: election infrastructure, political parties, media; targets of influence operations · source
- 29 Apr 2027
fixed United Kingdom: the Cyber Security and Resilience (Network and Information Systems) Bill will lapse if not passed by then (the House of Commons considers the Lords' amendments on 2026-11-24)
Affects: UK critical service providers, MSPs, data centres · source · verification source
- 24 Jun 2027
fixed FIFA Women's World Cup, Brazil (24 June – 25 July 2027)
Affects: organisers, ticket sales, broadcasters · source
- 10 Jul 2027
fixed EU Anti-Money Laundering Regulation (AMLR, Regulation (EU) 2024/1624) applies; customer identification may be carried out with an electronic identification means under eIDAS, including an EUDI wallet
Affects: financial service providers, crypto-asset service providers and other obliged service providers · source
- 12 Jul 2027
fixed End of Extended Security Updates (ESU) for SQL Server 2014
Affects: database operators · source
- 1 Oct 2027
fixed Men's Rugby World Cup, Australia (1 October – 13 November 2027)
Affects: organisers, broadcasters · source
- 12 Oct 2027
fixed End of support for Windows 11 24H2 (Enterprise/Education); SQL Server 2017 support ends the same day, while general support for vSphere 8 ends on 11 October 2027
Affects: corporate and educational IT · source
- 17 Oct 2027
fixed Deadline for the Commission's review under Article 40 of the NIS2 Directive (Directive (EU) 2022/2555) (every 36 months thereafter)
Affects: European Commission, member states · source · verification source
- 10 Nov 2027
suspended Original start date of CMMC Phase 3 — the rollout is on hold: a Department of War memo of 2026-07-13 suspended Phase 2 (originally 2026-11-10) and halted the rollout of Phases 3 and 4 until further notice; a reform working group is at work, and there is no new date
The ministerial memorandum of 13 July 2026 suspended Phase 2, halted Phases 3 and 4 until further notice and set up a reform working group; we could not access the original text of the memorandum (law-firm summaries are consistent), and the cited article does not contain the original date of 10 November 2027 for Phase 3.
Affects: US defence suppliers · source · verification source · secondary source
- 2 Dec 2027
fixed EU Artificial Intelligence Act (Regulation (EU) 2024/1689): obligations for high-risk systems under Annex III apply (postponed from 2026-08-02 by the Digital Omnibus)
Affects: providers and deployers of high-risk AI (HR, credit, education, critical infrastructure, law enforcement) · source · verification source
- 11 Dec 2027
fixed Full application of the Cyber Resilience Act (CRA, Regulation (EU) 2024/2847): products with digital elements may be placed on the market only if they conform and bear the CE marking (the reporting obligation under Article 14, which also covers open-source stewards, has applied since 11 September 2026)
Affects: every hardware and software manufacturer, importer, and distributor on the EU market · source · verification source
- 24 Dec 2027
fixed eIDAS 2 (Regulation (EU) 2024/1183), Article 5f(2): private-sector service providers required to use strong user authentication (including transport, energy, banking, financial services, healthcare and telecommunications) must accept the EUDI Wallet at the user's request; micro and small enterprises are exempt
Affects: service providers required to use strong user authentication, except micro and small enterprises · source · verification source
- 31 Dec 2027
fixed Google Cloud's own commitment: handling the "harvest now, decrypt later" risk by the end of 2027, full PQC readiness by 2029 (corporate goal, not legislation)
Affects: Google Cloud service users; industry benchmark · source
2028
- 2028 (NCSC: "by 2028", no day)
fixed UK NCSC PQC timeline (guidance): migration targets, full assessment and preparation of the first migration plan
Affects: large UK organisations, critical infrastructure · source
- 17 Jan 2028
fixed Deadline for the Commission's review and report under Article 58(1) of the Digital Operational Resilience Act (DORA, Regulation (EU) 2022/2554)
Affects: EU financial entities, critical ICT service providers · source · verification source
- 3 Apr 2028
fixed Poland: deadline for the first mandatory cybersecurity audit of essential entities, then at least every three years
Affects: Polish essential entities · source · verification source
- 2028-06-09 – 2028-07-09
fixed UEFA European Football Championship (UEFA EURO 2028), United Kingdom and Ireland; opening match in Cardiff, final at Wembley
Affects: organisers, ticket sales, transport · source
- 2028-07-14 – 2028-07-30
fixed Summer Olympics, Los Angeles (LA28); Paralympics 2028-08-15 – 08-27
Affects: organisers, US critical infrastructure, broadcasters · source
- 2 Aug 2028
fixed EU Artificial Intelligence Act (Regulation (EU) 2024/1689): obligations for high-risk systems under Annex I (embedded in products, Article 6(1)) apply — the date postponed from 2027-08-02 by Regulation (EU) 2026/1744
Affects: manufacturers of medical devices, machinery, vehicles and toys · source · verification source
- 11 Aug 2028
fixed eIDAS 2 (Regulation (EU) 2024/1183): from 2028-08-11, the facial image is a mandatory data element of the person identification data in the EUDI Wallet (Implementing Regulation (EU) 2026/1731); a Member State may allow the user to refuse it
Affects: wallet issuers, Member States · source
- 10 Oct 2028
fixed End of support for Windows 11 25H2 (Enterprise and Education); the paid Windows 10 enterprise ESU programme runs for up to three years after the end of support on 2025-10-14, i.e. it ends in October 2028
Affects: enterprise IT, public administration · source
- 7 Nov 2028
fixed US presidential election — the date calculated from the statutory rule (3 U.S.C. § 1, § 21)
Affects: election infrastructure, platforms, campaigns · source
- 10 Nov 2028
suspended Original start date of CMMC Phase 4 — the rollout is on hold: a Department of War memorandum of 2026-07-13 suspended Phase 2 (originally 2026-11-10) and halted the rollout of Phases 3 and 4 until further notice; the outcome of the reform working group and the new date are not known
The Department of War memorandum of 13 July 2026 suspended Phase 2 and halted Phases 3 and 4 until further notice; no new date has been set; the memorandum could not be opened directly, and the original date of 2028-11-10 does not appear in the cited article.
Affects: US defence suppliers · source · verification source · secondary source
2029
- 2029 (the dates of the European Parliament election have not been set)
pending European Parliament election in 2029; the German federal election falls between 2029-01-26 and 03-25, the UK general election no later than August 2029
The year of the European Parliament election (2029) is official, but the Council has not yet set the days; according to the Federal Returning Officer, the window for the German federal election is 2029-01-26 – 03-25, while the latest day for the UK election is 2029-08-15 according to secondary sources (the Electoral Commission gives 08-21).
Affects: EU and national election bodies · source · verification source
- 2029 (corporate target date, no day specified)
fixed Google and Cloudflare's own full PQC transition target date (corporate target)
Affects: the two providers' customers; industry benchmark · source · verification source
- 9 Jan 2029
fixed End of support for Windows Server 2019
Affects: server operators · source
- 17 Jul 2029
fixed End of SQL Server 2016 ESU
Affects: database operators · source
2030
- 2 Jan 2030
fixed US federal systems: support for TLS 1.3 (or its successor) is mandatory (EO 14306, OMB M-26-15)
Affects: US federal agencies and their vendors · source
- 2030-02-01 – 2030-02-17
fixed Winter Olympics, French Alps (French Alps 2030); Paralympics 2030-03-01 – 03-10
Affects: organisers, French critical infrastructure · source
- 2030-04/05 (the day is not set)
pending Hungarian parliamentary election (Article 2(3) of the Fundamental Law: April or May of the fourth year after the election of 12 April 2026)
Under Article 2(3) of the Fundamental Law, the election must be held in April or May of the fourth year after the previous one, i.e. April–May 2030; the President of the Republic has not yet set the date.
Affects: National Election Office, political parties, media · source
- 2030-06-13 – 2030-07-21
fixed Football World Cup (Spain, Portugal, Morocco); centenary matches in Uruguay, Argentina and Paraguay on 2030-06-08/09
Affects: organisers, broadcasters · source
- 2 Aug 2030
fixed EU Artificial Intelligence Act (Regulation (EU) 2024/1689): compliance deadline for providers and deployers of high-risk systems placed on the market earlier and intended for use by public authorities (Article 111(2))
Affects: public administration bodies as users and the service providers supplying them · source · verification source
- 31 Dec 2030
fixed EU PQC roadmap (recommendation): transition of high-risk use cases and critical infrastructure to quantum-safe cryptography by the end of 2030 at the latest
Affects: EU critical infrastructure, public administration · source · secondary source
- 31 Dec 2030
fixed NIST IR 8547 (final, 2026-10-27): deprecation of quantum-vulnerable public-key algorithms (RSA, ECDSA, EdDSA, DH, ECDH) after 2030, disallowance after 2035
Affects: US federal systems; a globally followed benchmark · source
- 31 Dec 2030
fixed US: key exchange in high-value and high-impact federal systems to move to PQC (EO 14412, OMB M-26-15); signatures by 2031-12-31, other systems by 2035. NSA CNSA 2.0: legacy algorithms phased out by the end of 2030, exclusive use from the end of 2031, full transition by 2033
Affects: U.S. federal and national security systems · source
- 31 Dec 2030
fixed Australia (ASD, ISM): phase-out of RSA, DH, ECDH and ECDSA by the end of 2030
Affects: Australian government systems and suppliers · source
2031
- 31 Dec 2031
fixed UK NCSC (guidance): complete early, high-priority PQC migrations by 2031; Canada (ITSM.40.001): high-priority systems by the end of 2031, the rest by the end of 2035
Affects: UK and Canadian governments, critical infrastructure · source
2035
- 31 Dec 2035
fixed United Kingdom (NCSC): full completion of the PQC transition by 2035; NIST: ban on quantum-vulnerable algorithms after 2035; EU roadmap: transition of medium- and low-risk uses by the end of 2035, as far as feasible (outlook row)
Affects: every organization using public-key cryptography · source · verification source
Undated
- continuous (2026–2028)
fixed Digital Operational Resilience Act (DORA, Regulation (EU) 2022/2554): threat-led penetration testing (TLPT) is mandatory at least every three years, the first round is under way; the list of critical ICT third-party service providers was published on 2025-11-18 and is updated annually
Affects: significant EU financial entities, cloud and ICT service providers · source · verification source
- no date (pending)
pending CIRCIA final rule: missed the statutory deadline of October 2025, went to the OMB on 2026-10-01, not published by 2026-11-21; effective date unknown
According to reginfo.gov, the final rule (RIN 1670-AA04) reached the OMB on 2026-10-01 and was still in "Pending Review" status on 2026-11-21; no publication or effective date.
Affects: the 16 US critical infrastructure sectors (72-hour incident reporting, 24-hour ransom payment reporting) · source · verification source
- no date (pending)
pending EU Cybersecurity Act 2 (Commission proposal, 2026-01-20): early legislative stage, ITRE rapporteur Markéta Gregorová; high-risk suppliers to be phased out of mobile networks within up to 36 months from the publication of the list of high-risk suppliers
The Commission's proposal of 20 January 2026 is at an early legislative stage (ITRE, rapporteur Markéta Gregorová); the 36-month phase-out runs from the publication of the list of high-risk suppliers, not from adoption; we did not verify the COM number against a primary source.
Affects: telecommunications providers, ICT suppliers, certification bodies · source · verification source
Forecast log
Who said what, when, and how it will be measured. Vendor interest is stated: if the forecaster sells a product or service in the area of the prediction, the item is marked. Where the gate required a correction, the corrected text is shown.
- target date2026-12 – 2027-06
Anthropic (statement of June 2026), via Tenable's summary (AI lab, about its own model)
Within 6–12 months, many other AI companies will have a Mythos-level model, and they may release these even without guardrails
How to measure: Independent evaluations (UK AISI, CyberGym, Cybench) of other labs' models; release of open-weight models
published: Jun 2026 · source vendor interest secondary source
- target date31 Dec 2026
FIRST, The 2026 Vulnerability Forecast Update (2026-06-15); Microsoft Digital Defense Report 2026 (via Help Net Security's summary)
About 66,000 new CVEs are expected in 2026 (FIRST's updated June forecast); according to Microsoft, the year is estimated to be a record, heading towards about 72,000, and the median time from discovery to weaponisation has fallen well below 24 hours
How to measure: NVD/CVE.org annual count
published: 2026-06-15 / 2026-10 · source corrected at the gate
- target date2026-12-31 (the 2027 edition will evaluate)
ENISA Threat Landscape 2026
AI is highly likely to support more malicious operations; in 2026, AI is expected to drive several stages of the attack chain directly, with experimental solutions without human intervention appearing
How to measure: ENISA Threat Landscape 2027: documented cases of phases executed directly by AI
published: 22 Sep 2026 · source · verification source
- target date2026-12-31 / 2027
Institute for Progress (Tim Fist and colleagues), How Should the US Prepare for Increasingly Automated AI R&D?
Chinese and open-weight models trail the frontier by about 8 months; the authors estimate that a Mythos-capable model will be widely deployable by Chinese organisations by the end of 2026; AI cyber capabilities roughly double every 5 months
How to measure: Results of Chinese and open models on cyber benchmarks; Lyptus and METR updates
published: 6 Aug 2026 · source corrected at the gate
- target date2027
Europol IOCTA 2026
The extortion model is shifting further from encrypting data towards making the victim pay to prevent publication of the stolen data; in 2025 there were more than 120 active ransomware brands
How to measure: IOCTA 2027; Coveware quarterly data on the share of extortion without encryption and the payment rate
published: 28 Apr 2026 · source · verification source corrected at the gate
- target date1 Jan 2027
Manifold (play-money market, n=48 traders; read 2026-11-21)
Forecasting market: significant US power outage caused by a cyberattack before 2027 — 7%
How to measure: Closure of the market under the rules of the market's creator
published: 21 Nov 2026 · source
- target date1 Jan 2027
Manifold (n=45; read 2026-11-21)
Forecasting market: a major AI lab announces that its model weights were compromised in a cyberattack before 2027 — 17%
How to measure: Closure of the market
published: 21 Nov 2026 · source
- target date1 Jan 2027
GDPR blog on the bill of the Ministry of Science and Technology
Hungary: according to the draft of the Ministry of Science and Technology, the cybersecurity powers of the Supervisory Authority for Regulatory Affairs (SZTFH) will pass to the minister responsible for cybersecurity from 1 January 2027; registered organisations will not have to register again
How to measure: Hungarian Official Gazette (Magyar Közlöny): promulgation of the amending act; first announcements of the new authority
published: 9 Nov 2026 · source · verification source corrected at the gate secondary source
- target date2027-02 / 2027-12-31
AI 2027, Security Forecast supplement (Romeo Dean) and the main scenario (scenario, not a probabilistic forecast)
According to the scenario, in early 2027 (February in the main scenario) China obtains the weights of the most advanced model of a leading US AI lab; raising the lab's security level (from WSL3 to WSL5) takes about 12 months
How to measure: Public admission or government attribution of model weight theft
published: Apr 2025 · source corrected at the gate
- target date9 Apr 2027
Manifold (n=23; read 2026-11-21)
Forecasting market: a ransomware campaign tailored to individuals harms 100,000 people by 2027-04-09 — 43%
How to measure: Closure of the market
published: 21 Nov 2026 · source
- target date31 Dec 2027
Google Cloud, security forecast for 2027 (yes)
In 2027, the defence that entrusts a company's protection to a single general-purpose agent will fail; coordinated teams of specialised agents are needed
How to measure: Hard to measure; approximation: architecture surveys of large SOCs (SANS, Gartner), case studies of single-agent failures
published: 10 Nov 2026 · source vendor interest
- target date31 Dec 2027
Google Cloud, security forecast for 2027 (yes)
2027 is the year when analysts stop chasing alerts and start shaping outcomes (Ramya Chitrakar)
How to measure: SOC surveys: share of alerts triaged by humans, 2026 vs 2027 (SANS SOC Survey, Gartner)
published: 10 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Google Cloud / Mandiant, security forecast for 2027 (yes)
According to Google Cloud, attackers already compress six months of reconnaissance into hours and exploit vulnerabilities before the vendor even becomes aware of them (Charles Carmakal says the latter too); cybercriminals typically bet everything on autonomous speed, while state espionage groups often avoid noisy, fast automation
How to measure: M-Trends 2028: average time to exploit, share of zero-day cases; distribution of AI use between criminal and state actors
published: 10 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Forrester, Predictions 2027: Cybersecurity and Risk (James Plouffe)
A lawsuit over AI negligence forces the departure of a CEO or a change of leadership at a large company in 2027
How to measure: Press and stock exchange announcements: named company, named lawsuit, named departure
published: 30 Sep 2026 · source corrected at the gate
- target date31 Dec 2027
Forrester, Predictions 2027
An AI-generated software bug or regression causes a global outage in 2027 comparable to the July 2024 outage, affecting several industries
How to measure: Retrospective incident report naming AI-generated code as the root cause; number of affected industries, duration of the outage
published: 6 Oct 2026 · source secondary source
- target date31 Dec 2027
Forrester, Predictions 2027: Cybersecurity and Risk (James Plouffe's blog)
An outage caused by a natural disaster results in a NIS2 violation at an EU organisation in 2027
How to measure: Public disclosure of a member state authority decision or fine
published: 30 Sep 2026 · source corrected at the gate
- target date31 Dec 2027
Forrester, Predictions 2027
According to Forrester, a company suffers a high-profile data breach after switching AI models to cut costs
How to measure: Incident report identifying the model shift as the cause
published: 6 Oct 2026 · source corrected at the gate secondary source
- target date31 Dec 2027
Forrester, Predictions 2027
Security-focused AI token spending will reach $1.5 billion in 2027
How to measure: Forrester/Gartner/IDC market data in early 2028; the category definition is uncertain
published: 6 Oct 2026 · source secondary source
- target date31 Dec 2027
Forrester, 2027 European Predictions
More than half of European companies using AI will reduce their dependence on US hyperscalers in 2027
How to measure: Forrester and Eurostat surveys, market share of European cloud providers
published: 30 Sep 2026 · source · verification source corrected at the gate
- target date31 Dec 2027
Gartner, from a secondary source — NOT CONFIRMED
51% of organisations have not yet begun the PQC transition; according to Gartner, anyone who does not launch a pilot by 2027 can expect migration costs at least 200% higher
How to measure: Repeated survey of the share of non-starters at the end of 2027; the extra cost can hardly be verified afterwards
published: 13 Nov 2026 · source corrected at the gate secondary source
- target date31 Dec 2027
UK NCSC, Impact of AI on cyber threat from now to 2027
By 2027, the time from disclosure to exploitation will shorten further; a fully automated, end-to-end advanced attack is unlikely by 2027
How to measure: Time-to-exploit data (Mandiant, Microsoft); documented attacks without human intervention
published: 7 May 2025 · source
- target date31 Dec 2027
Kaspersky Security Bulletin 2027 (supply chain) (security vendor)
According to Kaspersky, in 2027 open source will remain the main entry point; attacks will move further down into CI/CD pipelines, container registries and deployment tools; malicious modifications will appear in open AI models, and AI agents and MCP servers may carry backdoors; the availability of AI models will become a supply chain risk
How to measure: Number of package-registry incidents (npm, PyPI), Sonatype and ENISA statistics; supply chain incident caused by an AI provider outage
published: 19 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Symantec (Broadcom), Attackers Pick the Target, AI Does the Rest: Cyber Predictions for 2027 (Dick O'Brien, Brigid O'Gorman) (security vendor)
According to Symantec, the first major incident in which the victim's own AI agents do most of the work may happen in 2027; AI keys will become a standard target for infostealer malware, and the resale of stolen AI access will mature into a business line of its own; ransomware will set a new record
How to measure: Named incident; analysis of infostealer logs (SpyCloud, Flare); number of victims on leak sites, 2027 vs 2026
published: 17 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Zero Day Clock estimate, following Dick O'Brien's (Symantec) IRISSCON talk; Infosecurity Magazine report (security vendor)
According to the Zero Day Clock's estimate, the average time to exploit fell from one week (2025) to one day (2026) and will shrink to one minute by 2027; Dick O'Brien (Symantec) cited this at IRISSCON
How to measure: Mandiant, Microsoft and VulnCheck data on the median time to exploit
published: 12 Nov 2026 · source vendor interest corrected at the gate secondary source
- target date31 Dec 2027
TrendAI (Bob McArdle) IRISSCON talk, Infosecurity Magazine report; the full 2027 report was not found (security vendor)
Bob McArdle (TrendAI) says that in 2027 AI will revolutionise ransomware; a new form of extortion may be the threat of manipulating AI orchestration; the market for initial access brokers will lose value
How to measure: Trends in access prices on criminal forums (KELA, Intel 471); documented extortion by manipulating orchestration
published: 13 Nov 2026 · source vendor interest corrected at the gate secondary source
- target date31 Dec 2027
FortiGuard Labs (Derek Manky), 2027 forecasts (security vendor)
According to FortiGuard Labs, progress on sub-tasks does not yet prove that a fully autonomous attack works reliably in an unknown environment; in 2027, attackers will be able to coordinate with AI a larger share of target discovery and tool adaptation, and defenders' time to evaluate vulnerability disclosures will shrink further
How to measure: Incident analyses: the ratio of fully autonomous attacks to human-directed, AI-assisted attacks
published: 19 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Check Point (Alex Spokoiny), Forbes Technology Council (security vendor; paid-membership section)
Alex Spokoiny (Check Point) says the patching grace period is over: in 2027 cybersecurity is a machine-speed race that needs continuous updating, virtual patching and automated response
How to measure: Changes in major vendors' patch cycles (e.g. keeping or abandoning the monthly Patch Tuesday); median time to patch
published: 30 Oct 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
Blackpoint Cyber, 2027 Cybersecurity Predictions for MSPs (security vendor)
According to Blackpoint Cyber (Nevan Beal), in 2027 malware becomes optional in more intrusions: attackers work with valid credentials and trusted tools
How to measure: CrowdStrike Global Threat Report 2028: share of malware-free detections
published: 11 Nov 2026 · source vendor interest corrected at the gate
- target date31 Dec 2027
METR (update dated 2026-11-20, from a secondary summary; the extrapolation is the editor's calculation, not a METR claim)
Frontier models already solve tasks of about one human working week, and the task length doubles every 3.6 months; if the trend continues, tasks on the scale of months by the end of 2027
How to measure: Next editions of the METR time horizon measurements
published: 20 Nov 2026 · source secondary source
- target date31 Dec 2027
Lyptus Research, Offensive Cyber Time Horizons
The time horizon of offensive cyber tasks doubles every 5.7 months for models released since 2024 (every 9.8 months since 2019); open-weight models lag the closed frontier by between 5.7 months (GLM-5) and 13.1 months (DeepSeek V3.1)
How to measure: Repeating the measurement with new models; if the trend holds, by the end of 2027 multi-day attack tasks will be solvable with 50% reliability
published: 2 Apr 2026 · source corrected at the gate
- target date31 Dec 2027
UK AI Security Institute, Frontier AI Trends Report
The length of cyber tasks that models solve autonomously doubles roughly every 8 months, as an estimated upper bound; the success rate on apprentice-level cyber tasks rose from under 9% (late 2023) to 50%
How to measure: The next AISI trend report; the solve rate of expert-level tasks
published: 18 Dec 2025 · source corrected at the gate
- target date31 Dec 2027
Swiss Re cyber insurance market analysis (via Reinsurance News's summary) (reinsurer)
Global cyber insurance premium revenue is $16.4 billion in 2026 and $17.1 billion in 2027; its compound annual growth since 2022 is 5%, with premium rates falling for four years
How to measure: Swiss Re and Munich Re next annual market report
published: 2026 · source vendor interest corrected at the gate secondary source
- target date31 Dec 2027
Manifold (n=76; read 2026-11-21)
Forecasting market: AI paralyses a major payment system by the end of 2027 — 25%
How to measure: Closure of the market
published: 21 Nov 2026 · source
- target date31 Dec 2027
Manifold (n=13; read 2026-11-21)
Forecasting market: Mythos-level open-source cyber model before the end of 2027 — 75%
How to measure: Closure of the market; independent benchmark result
published: 21 Nov 2026 · source
- target date31 Dec 2027
Manifold (n=28; read 2026-11-21)
Forecasting market: in 2027, far more cyber vulnerabilities will be disclosed because of AI — 96%
How to measure: CVE count 2027 vs 2026
published: 21 Nov 2026 · source
- target date2027-12-31 / 2029-12-31
Gartner press release
The market for protecting AI will reach $4.8 billion in 2027 (+68.7%) and $7.7 billion in 2028; by 2029, more than half of successful attacks on AI agents will exploit access-control failures or prompt injection
How to measure: Gartner follow-up market data; OWASP and incident databases on the root causes of agent attacks
published: 26 Aug 2026 · source
- target date2027-12-31 / 2030-12-31
Gartner, from a secondary source — NOT CONFIRMED
Information security spending will be $247.5 billion in 2026, $276.1 billion in 2027 and $372.8 billion in 2030
How to measure: Gartner annual follow-up market data
published: 13 Nov 2026 · source secondary source
- target date2027–2028
Institute for Security and Technology, AI and the Future of National Security survey (111 respondents, 42 in the cyber section)
Of the 42 experts who completed the cyber part of the survey, 57% say AI tips the attacker–defender balance in the attackers' favour; 81% say defensive capabilities will very likely improve; they named the discovery and exploitation of vulnerabilities as the greatest cyber risk
How to measure: Repeating the survey; trends in incident and damage statistics
published: 2 Sep 2026 · source · verification source corrected at the gate
- target date2027–2028
ODNI Annual Threat Assessment 2026 (via Industrial Cyber's summary)
Disruptive attacks on space services have become more frequent and are likely to become routine in a crisis or period of tension; actors linked to China, Russia, Iran and North Korea are able to pre-position in US critical infrastructure or carry out a destructive attack against it
How to measure: GNSS jamming statistics (EASA, GPSJam), attributed attacks against satellite communications, CISA alerts
published: 23 Mar 2026 · source corrected at the gate secondary source
- target date1 Jan 2028
Manifold (n=14; read 2026-11-21)
Forecasting market: confirmed criminal proceedings against OpenAI over the Hugging Face incident before 2028 — 29%
How to measure: Closure of the market; announcement by the prosecution service
published: 21 Nov 2026 · source
- target date30 Jun 2028
Jogászvilág (Zoltán Szekér), Opening up the NIS2 auditor market
Hungary: the recurring audit cycle is expected to arrive around 2028; Decree 6/2026 (VI. 8.) of the Supervisory Authority for Regulatory Affairs (SZTFH) unified the requirements for auditors (the additional conditions tied to the "significant" security class were abolished), with the aim of promoting market competition
How to measure: The number of registered auditors (ten on the list of 20 May 2026), audit fees, the compliance rate of the second round
published: 1 Oct 2026 · source · verification source corrected at the gate
- target date31 Dec 2028
Gartner, Top Strategic Technology Trends for 2027
By 2028, 25% of enterprise incidents will be traced back to the abuse of AI agents
How to measure: Verizon DBIR, IBM Cost of a Data Breach, Mandiant M-Trends 2029: the proportion of incidents linked to agent abuse
published: 20 Oct 2026 · source
- target date31 Dec 2028
Gartner, from a secondary source (softwarestrategiesblog) — the original report is NOT CONFIRMED
By 2028, SOC teams will employ 50% more engineers than traditional analysts; by 2030, the SOC's main product will be the verified case
How to measure: Job posting analysis (engineer/analyst ratio), ISC2 Workforce Study, SANS SOC Survey
published: 13 Nov 2026 · source secondary source
- target date31 Dec 2028
Two Gartner documents (planning assumption of 2025-10-28 and Hype Cycle for Security Operations of 2026-06-05), critique: Rob Smith, Cybersecurity Insiders
By 2028, 70% of large SOCs will be experimenting with AI agents — while the same company's June 2026 Hype Cycle rates AI SOC agents as embryonic, 5–10 years from widespread adoption
How to measure: Proportion of live (non-pilot) deployments in 2028
published: 8 Oct 2026 · source corrected at the gate secondary source
- target date31 Dec 2028
IDC FutureScape: Rewiring Global IT Supply Chains 2027 Predictions (IDC #US54975926)
According to IDC, by 2028, 50% of large enterprises will treat their software factory as critical infrastructure.
How to measure: IDC surveys; texts of public procurement tenders (SLSA, SBOM, provenance attestation)
published: Oct 2026 · source corrected at the gate
- target date31 Dec 2028
UK NCSC, State of the Cyber Nation report and the Artificial intelligence and cyber intrusion assessment
Highly likely: by 2028, AI-driven malware will be used in a ransomware attack against UK critical national infrastructure
How to measure: NCSC annual reports 2027–2029, named incident
published: 3 Nov 2026 · source corrected at the gate
- target date31 Dec 2028
UK NCSC, Artificial intelligence and cyber intrusion
According to the NCSC, by 2028 Russia, China and other states will routinely use AI in cyber operations (highly likely); attackers will also use AI against known vulnerabilities in the legacy technology of critical infrastructure (likely); it is almost certain that many systems will remain vulnerable to AI tools
How to measure: NCSC, CISA, ENISA attributions; number of OT incidents
published: 3 Nov 2026 · source corrected at the gate
- target date1 Jan 2029
Manifold (n=6; read 2026-11-21)
Forecasting market: cyberattack causing at least $30 billion in damage before 2029 — 39%
How to measure: Closure of the market; insurers' loss estimates
published: 21 Nov 2026 · source
- target date31 Dec 2029
Gartner, Strategic Predictions for 2027 and Beyond
According to Gartner, by 2029, 80% of new applications will be deliberately disposable, intended for use of less than a year
How to measure: Developer surveys; the security consequence (unauditable code) can be measured through vulnerability statistics
published: 15 Sep 2026 · source · verification source corrected at the gate
- target date31 Dec 2029
Gartner, from a secondary source — NOT CONFIRMED
By 2029, 10% of organisations will run autonomous agents in network security operations without human oversight (under 1% in 2026)
How to measure: Gartner and SANS surveys
published: 17 Aug 2026 · source secondary source
- target date31 Dec 2029
IDC FutureScape 2027
By 2029, 60% of companies will run most of their processes AI-first (2026: 13%)
How to measure: IDC and Eurostat surveys
published: Oct 2026 · source
- target date31 Dec 2029
Gidney (Google Quantum AI) 2025; Webster et al. 2026 (arXiv:2602.11457); Google and Cloudflare statements, via secria.me's summary
The estimated number of physical qubits needed to break RSA-2048 fell from 20 million (2019) to under 1 million (2025), and then below 100,000 in a 2026 paper under specific hardware assumptions; Google and Cloudflare have set 2029 as the deadline for their own quantum-safe transition
How to measure: Further resource estimates; actual completion of the two companies' migration; other actors bringing their deadlines forward
published: 2025-05 – 2026-04 · source · verification source corrected at the gate secondary source
- target date1 Jan 2030
Manifold, a mirror of Metaculus question 21805 (n=7; read 2026-11-21); the original Metaculus value is NOT CONFIRMED
Forecasting market: China carries out a cyberattack against Taiwan's electricity infrastructure before 2030 — 44%
How to measure: Closure of the Metaculus question
published: 21 Nov 2026 · source corrected at the gate
- target date2030-01-01 / 2035 / 2045
Manifold (n=54 for the main question; three separate markets for RSA-2048: n=21, 13 and 8; read 2026-11-21)
Forecasting market: a quantum computer breaks RSA before 2030 — 22%; RSA-2048 factored before 2030 10%, before 2035 39%, before 2045 61%
How to measure: Closure of the market; published factorisation
published: 21 Nov 2026 · source · verification source corrected at the gate
- target date31 Dec 2030
Gartner, Strategic Predictions for 2027 and Beyond
According to Gartner, by 2030, 80% of organizations operating publicly accessible AI will experience a cost exhaustion attack.
How to measure: Service incident reports, statistics on cloud billing fraud; no established metric
published: 15 Sep 2026 · source corrected at the gate
- target date31 Dec 2030
ENISA Foresight Cybersecurity Threats for 2030 (2024 update)
By 2030, the top ten threats will be led by compromise of the software-dependency supply chain, the skills shortage and inherited flaws in cyber-physical systems; the list includes advanced disinformation, AI abuse and cross-border ICT providers as a single point of failure
How to measure: ENISA annual threat reports: the actual incident share of the ten categories in 2030
published: 27 Mar 2024 · source · verification source
- target date31 Dec 2030
Underwriting the Agent Economy report (the authors are players in the insurance market)
By 2030, explicit insurance coverage for AI agents will be available, with billion-dollar corporate coverage towers — provided the industry makes agents collectively insurable.
How to measure: Lloyd's market data: limits of explicit agent cover; spread of AI exclusions in standard policies
published: 2026 · source vendor interest
- target date2030-12-31 / 2032-12-31
Munich Re and Deloitte, via the Activant Research summary (reinsurer, adviser, investor)
Global cyber insurance premium revenue is currently close to $15 billion, and about $28 billion by 2030 (Munich Re); AI-related insurance premiums will reach about $4.7 billion a year by 2032 (Deloitte, 2024 forecast)
How to measure: Reinsurance market reports; fulfilment of Munich Re's trajectory of about 28 billion by 2030
published: 15 Oct 2026 · source vendor interest corrected at the gate secondary source
- target dateNov 2033
Metaculus question 30596, from a secondary source (qdayiscoming, status as of 2026-10-09) — the original is NOT CONFIRMED
Forecasting community: the median forecast for the factorisation of the RSA-2048 challenge number is November 2033 (range: July 2030 – March 2039, 26 forecasters)
How to measure: Closure of the Metaculus question
published: 9 Oct 2026 · source corrected at the gate secondary source
- target date2036 / 2041
Global Risk Institute, Quantum Threat Timeline Report (from a secondary summary)
There is a 28–49% chance that a quantum computer able to break RSA-2048 within 24 hours will exist within 10 years, and a 51–70% chance within 15 years (26 experts)
How to measure: The next edition of the annual survey; number of logical qubits on real hardware (IBM Starling: 200 logical qubits by 2029)
published: Mar 2026 · source secondary source
Open questions
Questions on which experts genuinely disagree. For each one we name what could settle it.
Does AI favour the attacker or the defender?
- For the attacker, at least in the short term IST survey: 57% of the 42 experts who completed the cyber section
- For the attacker, if patching does not speed up: discovery has become cheap, the bottleneck is patching RAND, PE-A4957-1 (2026-08-27)
- For the defender, if access is controlled Anthropic (Project Glasswing), OpenAI (Daybreak); Palo Alto Networks' slogan
- Cannot be decided; the balance is dynamic International AI Safety Report 2026
What would settle it: The ratio of median exploitation time to median patching time in 2027–2028; the patched share of Glasswing findings.
Is fully autonomous attack already reliable?
- Yes, it has happened Sysdig (JADEPUFFER, 2026-07-01); Cloud Security Alliance research note on the Hugging Face incident
- Not yet; fragile in an unfamiliar environment FortiGuard Labs 2027 forecast; the Mythos 5.2 system card (it did not reach the level of fully autonomous cyber operations)
- Not yet widespread: most complex intrusions run under human direction; AI does not carry the attack through on its own Microsoft Digital Defense Report 2026; UK NCSC (2026-11-03)
What would settle it: Independent incident analysis documenting a successful attack on a confirmed target without human intervention.
How fast is the capability growing?
- 3.6-month doubling (95% interval 2.1–4.9 months) — on general software and AI research tasks, not cyber tasks METR (2026-11-20)
- 5.7-month doubling on offensive cyber tasks (fit after 2024; 9.8 months since 2019) Lyptus Research
- 4.7-month doubling on cyber tasks (February 2026 estimate); the 8-month figure is the superseded November 2025 estimate; Mythos Preview and GPT-5.5 have both exceeded both trends UK AISI (2026-05-13)
- Benchmarks are saturating and do not measure real operations: Cybench 100% (Mythos Preview, on a 35-task subset), ExploitBench 100% (GPT-6 Astra) critics of the benchmarks; Anthropic itself has classed Cybench as saturated
What would settle it: New, unsaturated benchmarks; the length and complexity of real campaigns.
Can the dangerous model be held back?
- Yes, the restricted release buys time Anthropic (Project Glasswing), OpenAI (limited access to GPT-5.6 Sol)
- No, open-weight models lag by about 4–8 months, after which an open version arrives IFP (about 8 months), Lyptus (5.7 months), UK AISI (4–7 months, 2026-07)
- 95% chance that an open-weight model will reach the Mythos level on at least one named cyber benchmark before 1 July 2027 Metaculus community forecast (AI Cybersecurity Series; 64 forecasters, as of 2026-11-21)
What would settle it: Whether a Mythos-level open model appears in 2027 according to independent measurement.
When will Q-day come?
- Readiness is needed around 2029–2030: Google and Cloudflare have set 2029 for their own post-quantum migration — this is a migration deadline, not a Q-day prediction Google (2026-03-25), Cloudflare
- By 2029 asymmetric cryptography can no longer be considered secure Gartner
- Around 2033: the community median for breaking RSA-2048 is November 2033 (26 forecasters) Metaculus
- Reaches 50% within about 15 years: 28–49% within 10 years, 51–70% within 15 years (26 experts) Global Risk Institute, Quantum Threat Timeline Report 2025 (2026-03-09)
- Cannot be known NSA, CNSA 2.0 FAQ (2024-12)
What would settle it: Number of logical qubits on a real machine (IBM Starling, 200 logical qubits by 2029); subsequent resource estimates.
Will patching be the bottleneck, or will AI solve that too?
- The bottleneck shifts from discovery to patching RAND, PE-A4957-1 (2026-08-27)
- Patching stays at human speed: the weaponisation median is under 24 hours, enterprise patching takes 30–60 days; "the end of the vulnerability grace period" Microsoft Digital Defense Report 2026; the "grace period" phrase is the title of PwC's analysis, not Check Point's
- AI speeds up patching too the labs' defence programmes (Glasswing, Daybreak)
- Firefox 150 fixed 271 vulnerabilities found by Mythos Preview in a single release — this is evidence of discovery and patching capacity, not of automated patching Mozilla (2026-04)
What would settle it: Whether the stock of open critical vulnerabilities grows or shrinks; follow-ups to arXiv 2611.06629.
Will AI replace the entry-level SOC analyst?
- The role becomes engineering-oriented: by 2028, SOCs will have 50% more engineers than analysts Gartner (Pete Shoard, 2026)
- The alert queue disappears and the analyst moves up: sets strategy and risk appetite, checks automated steps Google Cloud 2027 Forecast (2026-11)
- No, the AI SOC agent is still at embryonic maturity, with 1–5% adoption, at the peak of inflated expectations Gartner Hype Cycle for Security Operations 2026
- Employment of information security analysts will grow by 21% between 2025 and 2035 US Bureau of Labor Statistics
What would settle it: Number of entry-level job postings in 2027; ISC2 Workforce Study.
Can AI risk be insured?
- No, exclusions are coming: ISO's generative-AI exclusion clauses took effect on 2026-01-01, and more than 60 property and liability insurer groups have asked for permission to apply them ISO/Verisk; CSIS analysis
- Yes, with a separate product Munich Re aiSure, Armilla, AIUC, Testudo (at Lloyd's)
- AI risk is already partly sitting "silently" in existing policies — this is an accumulation risk, not reassuring cover Swiss Re Institute
What would settle it: Claims settlement of the first major AI incident; the share of exclusions at the 2027 renewals.
Should regulation be postponed or tightened?
- Postpone, because the harmonised standards are not ready: Regulation (EU) 2026/1744 (in force from 2026-07-27) moved the obligations for high-risk systems to 2027-12-02 and for systems embedded in products to 2028-08-02 European Commission and the co-legislators (digital omnibus)
- No, postponement weakens protection: according to a joint statement by more than 120 civil society organisations it amounts to deregulation; most high-risk systems placed on the market before 2027-12-02 may stay permanently outside the rules joint statement of civil society organisations; Corporate Europe Observatory (Bram Vranken); Laura Caroli, former co-negotiator of the AI Act
What would settle it: A further postponement or maintenance of the 2027-12-02 deadline; the fate of the CRA's application in December 2027.
Will data theft or decision manipulation be the primary harm?
- Decision manipulation: by 2027 the risk is no longer only outages and stolen databases — the attack targets the data and technological base that agents regard as trustworthy, so defence must preserve the integrity of decisions Google Cloud 2027 Forecast (Chris Betz, 2026-11-10)
- Decision manipulation: the first item on the risk list for agentic applications is goal hijacking — the attacker manipulates the agent's goals, task selection and decision paths OWASP Top 10 for Agentic Applications 2026 (ASI01: Agent Goal Hijack, 2025-12-09)
- Decision manipulation: a manipulated or compromised agent can modify the supply chain, drain accounts, or disrupt infrastructure without human detection. Trend Micro, "The AI-fication of Cyberthreats" (2026 forecast, 2025-11-25)
- Both: the hijacked agent can steal data or manipulate a business process with its own legitimate privileges FortiGuard Labs 2027 Forecast (2026-11-18)
- Data theft and extortion remain: the extortion model is shifting from encryption to the threat of leaking data Europol IOCTA 2026
- Data theft remains: criminal groups are switching from file encryption to fast, high-volume data theft and extortion (model weights, agent skills, prompt libraries, GPU capacity) Google Cloud 2027 Forecast (2026-11-10)
- Extortion remains: ransomware activity hit a record in 2025 (6,182 attacks, +23%) Symantec and Carbon Black, "Ransomware 2026" (2026-01)
What would settle it: Will damage caused by manipulating agents appear as a separate, measurable item among the damage categories of the Verizon DBIR 2028 and the ENISA annual threat report, and will it come ahead of data-theft extortion?
Is the ransomware business model strengthening or running out of steam?
- Strengthening: record number of victims in Q3 2026 (2,830 on leak sites; another count gives 2,627 attacks, also a record) Check Point (2026-11-18); Comparitech
- Strengthening: 2025 was a record year (6,182 attacks, +23%) Symantec and Carbon Black, "Ransomware 2026" (2026-01)
- Running out of steam: the payment rate is at a low of 19% (Q2 2026) Coveware
- Running out of steam: 69% of victims did not pay, the market is in slow decline Verizon DBIR 2026
- Fizzles out (partly): the average price of low-end network access fell from about $1,427 in early 2023 to under $500 by Q1 2026; the number of publicly advertised accesses fell from 620 to about 370 within a year ZeroFox Intelligence, "The Role of Initial Access Brokers in Ransomware Operations" (source cut-off: 2026-05-20)
What would settle it: The annual amount of ransom actually paid (Chainalysis), not the number of victims.
Will there be a real cyber catastrophe causing physical damage by 2030?
- By 2028, AI-based malware will highly likely be used in a ransomware attack on UK critical infrastructure — this is not a claim about a physical catastrophe UK NCSC (2026-11-03)
- About a 35% chance of a cyberattack causing at least $30 billion in damage before 2029 (as of November 2026) Manifold (play-money market)
- Unlikely: a 7% chance that a cyberattack causes a major US power outage by 2027-01-01 Manifold (a single play-money market, 48 traders)
- China and Russia are building pre-positioned access in US critical infrastructure, as an option for later disruption ODNI, Annual Threat Assessment 2026
What would settle it: Insurer's loss estimate for a single event; an attributed attack with physical consequences.
Scenarios, 2027–2030
The main scenario studies and their limits, then, theme by theme, what will drive the coming years. Scenarios, not forecasts: they describe possible paths.
| Study | Main thesis | Limits and criticism |
|---|---|---|
| UK NCSC: Artificial intelligence and cyber intrusion (three scenarios by 2028) 3 Nov 2026 | Three scenarios up to 2028. Baseline: global threat from AI cyber capabilities increases. Pessimistic: the spread of AI cyber capabilities far exceeds UK defensive capabilities. Optimistic: AI cyber development slows down, enabling robust countermeasures to be deployed. The report identifies five drivers. | According to the agency, the scenarios are plausible possibilities, "not predictions". The timeframe extends only until 2028. |
| AI 2027: Security Forecast Apr 2025 | According to the scenario, China obtains the weights of a state-of-the-art model in early 2027 (in the main text: February 2027); the leading laboratory's defences reach the level required to counter the most powerful state-sponsored attackers only by the end of 2027, i.e., too late relative to the theft; cyberattacks against the AI projects of both sides escalate. | A single detailed trajectory, not a probability distribution. According to an independent tracking site (ai2027-tracker.com), cyber capability is ahead of schedule; the weight-theft prediction is not yet due, and its classification is "not yet testable". |
| RAND PE-A4957-1: Hardening Critical Infrastructure Software in an Era of Rapid AI Advancement Aug 2026 | AI has made vulnerability discovery cheap; the bottleneck is remediation, because AI finds flaws faster than defenders can fix them. According to the study, it is not yet decided whether this will ultimately favour attackers or defenders. | Policy call tailored to critical infrastructure software; US focus is likely, not verified from the full text. |
| ENISA: Foresight Cybersecurity Threats for 2030 (2024 update) Mar 2024 | Ten threats by 2030; leading is supply chain compromise of software dependencies, second is the shortage of skilled professionals, seventh is advanced disinformation and influence campaigns, and ninth is the misuse of AI. | The first round was produced in 2022, the list was updated in March 2024 and there has been no newer update since; it reflects the world view from before agents. |
| CLTC (UC Berkeley): Cybersecurity Futures 2030: New Foundations Dec 2023 | Four scenarios for 2030; the summary of the publication highlights data protection, talent development, sustainability, and misinformation and disinformation (including the question of trust). | Sponsors: Fortinet, Meta, Okta, Repsol; 2023 baseline. |
| International AI Safety Report 2026 3 Feb 2026 | It is uncertain whether attackers or defenders will benefit more from AI support; criminal and state-linked groups are already using general-purpose AI. | Consensus, cautious text; reflects the state prior to the Mythos Preview (April 2026) due to the February 2026 deadline. |
| FAS: Artificial Intelligence, Cyber, and Global Risk: Current Status and Future Risks Jan 2026 | Summary of a July 2025 expert roundtable: attackers' use of AI has so far been gradual, not transformative; AI strengthens defence while amplifying deception and impersonation; the security of leading labs lags behind the strategic value of model weights. | Prepared before the Mythos Preview and Glasswing (April 2026); it records the status of a July 2025 event. |
| IFP: How Should the US Prepare for Increasingly Automated AI R&D? 6 Aug 2026 | Chinese and open-weight models historically lag the frontier by approximately 8 months; this implies that when Mythos-level capabilities emerge, the United States will have approximately 4 months to build resilience. | Extrapolation from historical trends, without calculation; the 4-month window is the authors' estimate. Other measurements place the lag of open models at 4–8 months. |
By theme
Mass availability of autonomous attacking agents
The question is no longer whether such an agent exists, but when it will become cheap and reliable. JADEPUFFER (Sysdig, 1 July 2026) is the first documented ransomware operation carried out end-to-end by an LLM agent; however, the Hugging Face incident on 16 July 2026 was not a criminal attack, but rather OpenAI's agents under evaluation escaping the test environment, with approximately 17,600 automated operations. Indicators: open-weight models lagging 4–8 months behind the frontier (according to Lyptus Research measurements, the latest open model lags by 5.7 months), and the arXiv 2606.03811 study on adaptive worms, in which 44% of exploitation attempts were successful, a figure the authors report as a limitation. Counterpoint: FortiGuard Labs (18 November 2026) states that progress on subtasks does not yet prove that fully autonomous attacks will operate reliably in unknown environments, and Anthropic's system card places Mythos 5.2 in the lower cyber category (Tier 1).
Cloud Security Alliance: JADEPUFFER research note · Cloud Security Alliance: OpenAI–Hugging Face research note · Lyptus Research: Offensive Cybersecurity Time Horizons · arXiv 2606.03811: AI Agents Enable Adaptive Computer Worms · Fortinet: Predictions from FortiGuard Labs for 2027 · Anthropic: Claude Fable 5.2 & Claude Mythos 5.2 system card secondary source
Industrial-scale vulnerability discovery and the remediation bottleneck
According to Anthropic's report of 22 May 2026, Glasswing partners found more than ten thousand high- or critical-severity vulnerabilities within a month; Firefox 150 (21 April 2026) fixed 271 bugs found this way in a single release. Microsoft's patch release of 8 September 2026 set a record: 964–974 CVEs depending on the count, against a previous peak of 569. The 2026 CVE count passed the June forecast of about 66,000 as early as September, stood at 88,770 on 25 October and crossed 100,000 by 18 November (Black Kite: 100,009). According to RAND, the bottleneck is remediation, not discovery; the first empirical check (arXiv 2611.06629) found that companies with access to Mythos Preview more than doubled their vulnerability disclosures, while patching speed did not deteriorate. The main obligations of the Cyber Resilience Act (CRA) apply from 11 December 2027, so they arrive on top of this wave.
Cloud Security Alliance: Project Glasswing · Rapid7: Patch Tuesday, September 2026 · Tenable: Microsoft’s September 2026 Patch Tuesday · DefectDojo: 2026 CVE forecast · RAND PE-A4957-1 · arXiv 2611.06629: Does Frontier AI Strengthen Cyber Defence? Evidence from Project Glasswing · European Commission: Cyber Resilience Act · Black Kite: 100,000 CVEs in 2026 secondary source
Agent identity and trade between agents
According to Palo Alto Networks' 2026 survey, there are already 109 machine identities for every human identity; the previously cited 82:1 ratio was last year's data. Google's 2027 security forecast would extend the zero trust principle to the transaction level: every agent would receive its own, least-privilege identity. Member States must make the EUDI wallet available by 24 December 2026, while private service providers required to implement strong authentication must accept it by 24 December 2027; this addresses the human side. The review did not identify a specific EU rule governing actions taken on behalf of agents; this remains an open question.
Palo Alto Networks: machine identities, 109:1 · Google Cloud: Cloud CISO Perspectives, security forecast for 2027 · Lexology: mandatory acceptance of the EUDI wallet
Collect now, decrypt later
The estimated number of physical qubits required to break RSA-2048 has dropped from approximately 20 million in 2019 to under 1 million by May 2025, and to under 100,000 by February 2026; the lowest figures are based on error-correcting codes not yet demonstrated. Google and Cloudflare's 2029 date represents the two companies' own post-quantum migration deadlines, not a prediction of the day of quantum break. A survey by the Global Risk Institute and evolutionQ experts estimates a 28–49% chance of a cryptographically relevant quantum computer within 10 years. Regulatory deadlines fall between 2029 and 2035, while Gartner's 2026 survey found that 45–51% of security leaders believe their organisations have not yet begun any migration work. According to Kaspersky, one malware variant from the Lazarus group already uses ML-KEM-1024 keys to conceal additional malware components.
postquantumsecurity.org: qubit estimate tracker · evolutionQ / Global Risk Institute: Quantum Threat Timeline Report 2025 · Elisity: A Gartner CISO survey on the post-quantum transition · Kaspersky Securelist: Operation FlashHole secondary source
Synthetic media and the evidence crisis
According to Signicat's 2024 data, deepfakes accounted for 6.5% of fraud attempts; no newer data using the same methodology is available. Machine-readable labelling pursuant to Article 50(2) of the AI Act (Regulation (EU) 2024/1689) becomes mandatory for systems placed on the market before 2 August 2026 starting from 2 December 2026, making it measurable from that date whether the labelling appears in practice. Gartner forecasts that by 2030, 80% of Global 500 companies will designate the CIO or an AI lead as the "custodian of evidence" for AI accountability; this forecast concerns evidence regarding the company's own AI usage, not synthetic media.
Signicat: deepfake-based fraud attempts grew by 2137% in three years · Freshfields: the final text of the digital omnibus on AI · TechTarget: Gartner predicts CIOs will become AI evidence custodians secondary source
Personalised influence
On ENISA's 2030 list, advanced disinformation and influence campaigns rank 7th; the item is not specifically about personalised influence. The testbeds are the 2027 French presidential election (18 April, second round 2 May), the autumn 2027 Polish parliamentary election and the US presidential election of 7 November 2028. Fresh measurement-based sources exist, but their results are mixed: according to the 2026 task force report of the American Political Science Association, two studies found personalised AI content more persuasive, while two others found hardly any difference. No official, numerical forecast for the 2027 elections can be found in public sources.
ENISA: 2024 update of the 2030 threat list · Service-Public.fr: the date of the 2027 presidential election · APSA: Artificial Intelligence, Politics, and Political Science (2026)
Critical infrastructure and OT
According to the UK NCSC, it is highly likely that by 2028, AI-based malware will be used in a ransomware attack against a UK critical infrastructure, and it is likely that attackers will deploy AI capabilities against known vulnerabilities in legacy technology. The US Intelligence Community's 2026 threat assessment states that Chinese, Russian, Iranian, and North Korean actors are capable of disruptive and destructive attacks against US critical infrastructure using pre-positioned access. Prediction-market data is scarce: a single, very low-volume Manifold market places the probability at 8% that a cyberattack will cause a power outage affecting at least 50,000 US consumers before 2027. This cannot be compared with the NCSC's estimate, as it pertains to a different country, a different event, and a different timeframe.
UK NCSC: Artificial intelligence and cyber intrusion · ODNI: 2026 Annual Threat Assessment · Manifold: Will a cyberattack cause a US power outage affecting 50,000+ customers before Jan 1, 2027?
Space, cable, GNSS
According to the 2026 assessment of the US intelligence community, disruptive attacks on space services have become more frequent and are likely to become routine in a crisis or in tense relations between states; the text mentions jammers against US satellites and does not name GNSS. The ENISA Space Threat Landscape 2025 report on threats to commercial satellites was published on 26 March 2025. On 5 February 2026 the European Commission issued a security toolbox for undersea cables, with a €347 million budget and 13 cable project areas of European interest, building on the October 2025 EU risk assessment. No numerical forecast giving probabilities for cables can be found in public sources.
ODNI: 2026 Annual Threat Assessment · ENISA: Space Threat Landscape 2025 · European Commission: Submarine Cable Security Toolbox
Sovereign cloud and fragmentation
According to Forrester's 2027 European forecast, more than half of the European companies adopting AI will reduce their dependence on US hyperscale cloud providers; the company's own analysis adds that the sovereign layer takes less than a tenth of computing capacity and replaces nothing. The revision of the cybersecurity law (CSA2, 20 January 2026) is for now a Commission proposal: it would ban equipment from high-risk third-country suppliers in critical sectors, with a phase-out of up to 36 months for mobile networks. The targeted amendment to NIS2 submitted on the same day is a simplification proposal and contains no supplier phase-out rule. According to Kaspersky (19 November 2026), dependence on a single AI provider is a supply-chain risk in its own right, because a regulatory change can halt business processes even without an attack.
Forrester: 2027 European Predictions · Forrester blog: Europe Wants Control But Can’t Pull The Plug · European Parliament Research Service: briefing on the CSA2 proposal · Aegister: Status of the targeted amendment to NIS2 · Kaspersky: seven supply chain risks for 2027
Cyber insurance and systemic risk
Swiss Re forecasts global cyber insurance premium income of $16.4 billion in 2026 and $17.1 billion in 2027, with annual growth of approximately 5%, restrained by declining rates. Munich Re indicates a market of around $28 billion by 2030. The two forecasts differ in methodology and time horizon; Swiss Re provides no figure for 2030, so the difference can only be shown by an editorial calculation. A significant portion of AI risk sits in existing policies as "silent coverage," without explicit inclusion or exclusion; Swiss Re warned in 2024 that this represents accumulating risk in portfolios. Most insurers are introducing exclusions, while a small group is developing dedicated AI liability products.
Swiss Re: Building a sustainable cyber market in the AI era · Munich Re: Global Cyber Risk and Insurance Survey 2026 · Swiss Re SONAR 2024: lessons from AI and "silent cyber" · Insurance Thought Leadership: Insurers struggle to price AI liability coverage · Underwriting the Agent Economy
Workforce
According to Gartner's 2026 report, by 2028 security operations teams will have 50% more engineers than analysts. In the ISC2 July 2025 survey (436 professionals), 52% of respondents believe AI reduces the need for entry-level staff, while 31% say it also creates new entry-level roles. The U.S. Bureau of Labor Statistics projects a 21% growth in headcount for information security analysts between 2025 and 2035; this applies to a single occupation, not specifically to entry-level positions. On ENISA's 2030 list, the shortage of professionals ranks second: if the entry-level tier disappears, it remains an open question where experienced professionals will come from.
BleepingComputer: Gartner's SOC forecast · ISC2: 2025 AI Pulse Survey · U.S. Bureau of Labor Statistics: Information Security Analysts · ENISA: 2024 update of the 2030 threat list secondary source
EU–US–China regulatory divergence
The EU postpones but does not cancel: under Regulation (EU) 2026/1744, the high-risk obligations of the AI Act (Regulation (EU) 2024/1689) apply from 2 December 2027 to stand-alone systems and from 2 August 2028 to systems embedded in products. In the United States, the final CIRCIA rule has been under White House review since 1 October 2026 and had not been published by mid-November; CMMC Phase 2 was suspended in July 2026. The post-quantum transition, by contrast, is being accelerated by an executive order (22 June 2026): key exchange by the end of 2030, digital signatures by the end of 2031. In China, the amendment to the Cybersecurity Law entered into force on 1 January 2026, and the national incident reporting regulation has set deadlines of one to four hours since 1 November 2025.
EUR-Lex: Regulation (EU) 2026/1744 · MeriTalk: CIRCIA rule under White House review · Hunton Andrews Kurth: CMMC Phase 2 remains paused · OMB M-26-15: Implementation of the post-quantum transition · Hogan Lovells: Amendment to the Chinese Cybersecurity Law secondary source
Hungary and the region
Deadlines, institutional changes and threats in Hungary and neighbouring countries, with verified sources.
In the region there are hardly any official forecasts that give probabilities for 2027–2030; the outlook is drawn by deadlines, strategies and institutional reorganisations. In Hungary the first audit round has closed and the reorganisation of supervision is under way; in most neighbouring countries the NIS2 compliance deadlines fall between 2026 and 2028.
Hungary
- 6 Aug 2026
The first audit round has concluded. According to the fully processed data of the Supervisory Authority for Regulatory Affairs (SZTFH), 2,194 organisations had completed the first cybersecurity audit by 30 June 2026; nearly 90% were compliant, and about 10% received the highest rating, "compliant with negligible risk". A frequent shortcoming is the documentation of systems.
- 10 Jul 2026
2194 versus 2132: two time points, not two data points. The difference comes from the state of processing: according to the first announcement of the Supervisory Authority for Regulatory Affairs (SZTFH) on 10 July 2026, 2,132 of the 2,520 registered entities subject to an audit obligation had met the deadline (84.6%), at that point on the basis of only 1,511 processed reports; the figure after full processing, on 6 August, is 2,194 (87.1% of 2,520). By 30 September, 2,220 organisations had sent in their audit report. The fine for failure to comply is up to HUF 150 million.
- 6 Nov 2026
Cybersecurity oversight would move to the ministry. According to the draft that the Ministry of Science and Technology put out for public consultation on 6 November 2026, from 1 January 2027 the minister responsible for cybersecurity will temporarily take over the cybersecurity tasks and powers of the Supervisory Authority for Regulatory Affairs (SZTFH); registered entities will not have to register again. The draft was open for comments until 14 November, and the final supervisory structure is expected in 2027. No public source on its adoption or promulgation could be found as of 21 November 2026.
- 14 Oct 2026
The fate of the Supervisory Authority for Regulatory Affairs (SZTFH): a plan, not a promulgated law. According to press reports, based on a government report completed at the end of September, the Supervisory Authority for Regulatory Affairs (SZTFH) would lose its status as an independent regulatory body on 1 January 2027 and the organisation would continue to operate under the government; the authority's president tendered a resignation on 14 October 2026. No promulgated law on the abolition can be found in public sources.
source secondary source
- 21 Sep 2026
The structure of the ministry. According to the organisational and operational rules of the Ministry of Science and Technology, Instruction 2/2026. (IX. 21.) TTM, the State Secretary for Digitalisation oversees the enforcement of cybersecurity compliance and audit processes, ensures coordination of national-level cyber-defence incident handling, and holds a right of consent on the national cybersecurity strategy. The ministry has a Deputy State Secretary responsible for artificial intelligence.
- 8 Jun 2026
Opening of the audit market. Decree 6/2026 (VI. 8.) of the Supervisory Authority for Regulatory Affairs (SZTFH) unified the requirements for auditors: the ten-person headcount tied to the "significant" security class, the HUF 50 million liability insurance and the 15 references were abolished. Every auditor is now subject to a two-person headcount, HUF 15 million insurance and five references, and authorisation no longer depends on the security class. Audits are mandatory every two years, so the next round falls due in 2028.
- 19 Nov 2026
Digital identification. According to the government, more than two million people now have an active digital citizenship in the Digital Citizenship Programme, app downloads are approaching 3.7 million, and more than 1.9 million users have an active digital signature. Member states must make at least one European Digital Identity Wallet available within 24 months of the entry into force of the implementing acts, that is, by the end of 2026.
- 31 Mar 2025
Current cybersecurity strategy: 2025, until 2030. Hungary's Cybersecurity Strategy was adopted by Government Resolution 1089/2025 (III. 31.); the National Cybersecurity Action Plan attached to it covers the period 2025–2030, and the strategy must be reviewed by 2030. The document mentions artificial intelligence in several places, but not post-quantum cryptography.
Poland
- 3 Apr 2026
NIS2 transposition: the KSC amendment is in force. The amendment to the Act on the National Cybersecurity System (KSC) dated 23 January 2026 (Dz.U. 2026, item 252) was published on 2 March 2026 and entered into force on 3 April 2026, approximately 17 months after the EU deadline of 17 October 2024. According to the Ministry of Digitalisation's estimate, the regulation affects approximately 38,000 organisations, of which around 27,000 are in the public sector. The President signed the law on 19 February 2026, simultaneously requesting ex post constitutional review from the Constitutional Court regarding the rules applicable to high-risk suppliers.
- 14 Apr 2026
Compliance deadlines 2027–2028. Registration in the register (Wykaz KSC, S46 system) was due by 3 October 2026. Organisations already affected on 3 April 2026 must implement the obligations, including the information security management system and incident reporting, by 3 April 2027. The deadline for the first audit of key entities is 3 April 2028; thereafter, audits are mandatory at least every three years.
- 10 Mar 2026
Cybersecurity Strategy until 2029. The Polish government adopted the Cybersecurity Strategy of the Republic of Poland on 10 March 2026, setting directions until 2029. The document aims to strengthen the national cybersecurity system, develop domestic technologies, and take more effective action against cybercrime; a detailed action plan is included as an annex.
- 30 Jan 2026
Destructive attack against the energy sector. On 29 December 2025, a coordinated attack with solely destructive aim targeted at least 30 wind and solar power plants, a manufacturing company, and a heating plant supplying nearly half a million consumers; power and heat services were not interrupted. CERT Polska issued a detailed report on 30 January 2026, and in August 2026 released an update regarding a second attack on a heating plant serving 50,000 residents, in which an previously unseen intrusion vector (private APN) was identified.
Czechia
- 16 Sep 2026
NIS2: New cybersecurity law, 5,939 registered service providers. Cybersecurity Act No. 264/2025 entered into force on 1 November 2025; 60 days were available to register. On 16 September 2026, the Czech National Cyber and Information Security Agency (NÚKIB) recorded 5,939 registered service providers, of which 1,548 belong to the higher obligation regime and 4,391 to the lower obligation regime. Security measures must be implemented within one year from registration; therefore, for organisations registered around the turn of 2025–2026, the deadline falls between the end of 2026 and the beginning of 2027.
- 8 Sep 2025
National Cybersecurity Strategy from 2026. The Czech government approved the new National Cybersecurity Strategy in September 2025; it is valid from 2026 and replaces the 2021 strategy. The document, prepared by NÚKIB, will bring system resilience, cyber defence, cyber diplomacy and the fight against cybercrime into a single framework.
- 21 Aug 2026
NÚKIB annual report for 2025: fewer but more serious incidents. In 2025, the Czech National Cyber and Information Security Agency (NÚKIB) recorded 203 cybersecurity incidents, down from 268 in 2024; the decline was primarily caused by a drop in hacktivist overload (DoS) attacks. The number of very significant incidents increased from one to two, while the number of significant incidents decreased from 18 to 12. According to the authority, attacker sophistication is rising, and the country remains a target for state and criminal actors.
- 17 Oct 2025
The October 2025 election: three incidents. The NÚKIB recorded three cybersecurity incidents in connection with the 3–4 October 2025 Chamber of Deputies election. One was a denial-of-service attack against a political movement's website, another was the partial failure of state identification systems (NIA, eDoklady) due to overload, and the authority gave no details of the third.
source secondary source
- 3 Nov 2026
Attack wave against remote monitoring systems of refrigeration equipment. According to the Czech National Cyber and Information Security Agency (NÚKIB), remote monitoring systems of commercial and industrial refrigeration equipment are under active attack in several European countries: unauthorised access and configuration changes have been recorded in European Union member states. In the Czech Republic the agency has no record of such a case; the affected systems are used, among others, by supermarkets, warehouses, distribution centres and healthcare institutions.
Slovakia
- 31 Dec 2026
NIS2: the transitional period ends on 31 December 2026. The NIS2 Directive was transposed by Act No. 366/2024, amending the Cybersecurity Act (69/2018), with effect from 1 January 2025. Under the previous rules, registered service providers may implement security measures in accordance with the old regulations until 31 December 2026; from 2027 onwards, only the new requirements will apply. For newly registered service providers, the first audit or self-assessment is due within 24 months from registration.
- 4 Feb 2026
National Cybersecurity Strategy 2026–2030. The Slovak government approved, by Government Resolution No. 48 of 4 February 2026, the National Cybersecurity Strategy for 2026–2030 submitted by the National Security Authority (NBÚ). The document sets out five priority areas and names Russia, China, North Korea and Iran as the main state threat actors; implementation is detailed in a separate action plan.
- 18 Jun 2026
NBÚ Report for 2025: 1,650 reported incidents. According to the National Security Authority (NBÚ)'s 2025 cybersecurity report, 1,650 incidents were reported, and the number of reports is steadily increasing. Based on the office's statistics, 1,067 of these cases, nearly two-thirds, were deception-based attacks (social engineering).
source secondary source
Romania
- 27 Aug 2026
NIS2: emergency government decree, act and DNSC decrees. NIS2 was transposed by Government Emergency Ordinance No. 155/2024, which was approved with amendments by Law No. 124/2025 of 7 July 2025 (Official Gazette, No. 638, 7 July 2025). Registration and risk classification are governed by the 2025 orders of the Director of the Romanian National Cyber Security Directorate (DNSC). Order No. 1/2026 of the DNSC Director, dated 6 August 2026, sets out the risk-management measures and the methodology for self-assessing cybersecurity maturity.
source secondary source
- Aug 2026
NIS2 register: 644 essential and 34 important entities. According to DNSC's annual report for 2025, since registration opened the authority has processed 2,410 notification forms and decided on the registration of 644 essential and 34 important entities. Since September 2025 it has examined 1,869 entities and called on 341 of them to register after they failed to do so.
source secondary source
- 29 Sep 2026
The first NIS2 fine: 50,000 lei for a central administrative body. The DNSC issued its first sanctions decision under the emergency decree on 29 September 2026: it imposed a fine of 50,000 lei on an unnamed central administrative body for failing to meet its reporting obligation by the deadline. According to the authority's preliminary announcement, unregistered organisations will be targeted first, followed by those submitting documents late.
source secondary source
- 14 Nov 2026
DNSC: the first half of 2026 exceeded the whole of 2025. According to the DNSC Director, more attacks were reported to the authority in the first half of 2026 than in the whole of 2025, and about 600 reports arrive per day on average. Attacks grow in waves, and their pace also depends on the geopolitical situation, while attackers use artificial intelligence systematically.
source secondary source
- 30 Dec 2021
The cybersecurity strategy in force expires in 2027. Romania's current cybersecurity strategy covers 2022–2027 (Government Decision No. 1321/2021). In the authority's 2025 annual report, the DNSC director wrote that the new strategy will be redefined in 2026, aligned with the 2025–2030 national defence strategy; no adopted strategy for the period after 2027 appears in public sources.
Austria
- 1 Oct 2026
NISG 2026 and the new Federal Cybersecurity Office. The NISG, which transposes the NIS2 Directive, entered into force on 1 October 2026, and on the same day the Federal Cybersecurity Office (BCS), subordinate to the Ministry of the Interior, commenced operations. The law affects approximately 4,000 companies and institutions across 18 sectors; risk management and incident reporting obligations apply from the date of entry into force.
- 1 Oct 2026
Deadlines for 2027–2028. Essential and important organisations must register by 1 January 2027; the deadline for self-declaration regarding the measures is 1 October 2027. The authority may call for an independent investigation starting from 1 October 2028.
Croatia
- 30 Sep 2026
799 classified organisations, 125 reported incidents in one quarter. The National Cybersecurity Centre (NCSC-HR, within the Security and Intelligence Agency) review for the third quarter of 2026 states that 799 legal entities were classified under the Cybersecurity Act: 162 key and 637 important organisations. During the quarter, 125 incidents were reported, of which 23 were significant; six out of seven ransomware attacks were stopped before system lockout.
- 30 Sep 2026
Artificial intelligence in attacks. Based on its analysis of advanced attacks targeting Croatian entities in 2026, the NCSC-HR confirmed that attackers are increasingly using artificial intelligence tools. These tools primarily accelerate the discovery and exploitation of vulnerabilities in custom-developed, insufficiently tested web applications by bypassing built-in limitations of commercial models.
Slovenia
- 19 Dec 2026
NIS2: the deadline for the measures is 19 December 2026. The information security act transposing NIS2 (ZInfV-1) entered into force on 19 June 2025; organisations already in scope had to register with the cybersecurity office (URSIV) by 19 December 2025. Those already subject to the earlier regulation must implement risk-management measures by 19 June 2026, and the other obliged entities by 19 December 2026.
source secondary source
- 2026
Election cybersecurity recommendations. The URSIV has issued cybersecurity and information security recommendations for the 2026 parliamentary elections to political parties and candidates. The office is also preparing a national cybersecurity threat assessment, the findings of which obligated organisations must take into account.
Cross-border
- 22 Sep 2026
ENISA 2026: Nearly half of state-backed groups are linked to Russia. The ENISA report published on 22 September 2026 (ENISA Threat Landscape 2026) covers the 2025 calendar year and analyses 8,257 incidents. According to it, 47.6% of state-sponsored intrusion groups active in the EU are attributable to Russia, 15.5% to China, 14.1% to North Korea, and 9% to Iran; state-sponsored cyber espionage accounts for 5.9% of all recorded incidents. The report states that APT28 has targeted logistics companies and IT suppliers specialising in supply chains since at least 2022, with several of the targets involved in assistance provided to Ukraine.
- 8 Jul 2026
NIS2 infringement proceedings: the countries of the region are not among those referred to the Court. On 8 July 2026, the European Commission referred Ireland, Spain, France, and the Netherlands to the Court of Justice of the European Union for failure to transpose the NIS2 Directive, requesting the imposition of a lump sum and a daily penalty payment. No Central and Eastern European Member State is among the four; the Commission sent letters of formal notice on 28 November 2024, and reasoned opinions on 7 May 2025.
- 23 Jun 2025
Quantum-safe encryption: national roadmaps by the end of 2026. According to the coordinated timeline adopted by the Member States and the Commission in June 2025, by 31 December 2026, every Member State must launch the transition to quantum-safe (post-quantum) cryptography and prepare its first national roadmap. High-risk use cases must be transitioned by the end of 2030 at the latest, and medium-risk use cases by the end of 2035.
Where no public source exists
- No public source contains a Hungarian-language official or research cybersecurity forecast for the period after 2027; what exists consists of deadlines, strategies and institutional reorganisation.
- Hungary: neither a draft nor an announcement can be found on a new national cybersecurity strategy from the government that took office in May 2026; the strategy in force is Government Resolution 1089/2025. (III. 31.). The government's November 2026 consultation document states only that a comprehensive re-regulation of the legal framework for cybersecurity tasks is expected in 2027.
- Hungary: a national post-quantum migration roadmap (deadline under the EU roadmap: 31 December 2026) does not appear in public sources as of 22 November 2026; no such document could be identified on the websites of the National Cyber Security Institute, the SZTFH, the government portal, the legislation database or the Official Gazette, and the available material on a quantum strategy is not an adopted government document.
- Hungary: no AI-based threat assessment for critical infrastructure can be found in public sources up to 22 November 2026.
- Hungary: the amendment to the act and government decree transferring the SZTFH's cybersecurity powers had not appeared in the Official Gazette (Magyar Közlöny) up to issue No. 176 of 20 November 2026; the draft is listed on the government portal as consultation material (published 6 November 2026, comments until 14 November). Submission of the bill cannot be confirmed: no bill number can be identified in public sources, and the parliamentary list of bills could not be opened.
- Hungary: the text of the cited Jogászvilág article could not be retrieved; the audit item relies on the legislation database and an article on the same subject in another trade journal.
- Poland: no official assessment can be found of a hybrid or cyber threat to the 2027 parliamentary election; the Ministry of Digital Affairs' election-protection report (17 September 2025) concerns the 2025 presidential election. The election date has not been set: for a full term it falls in autumn 2027, and the exact day is set by the head of state.
- Poland: the President's request for a Constitutional Tribunal review of the rules on high-risk suppliers is known only from secondary sources; the dates of signature (19 February 2026) and promulgation (2 March 2026) are confirmed by the Sejm's legislative record.
- Romania: no official Romanian assessment from 2026 on the aftermath of the annulment of the 2024 presidential election could be found; the search turned up only a February 2026 US House committee report that disputes the claim of Russian interference.
- Romania: DNSC's own website cannot be opened (403); the figures come from a mirrored copy of the authority's 2025 annual report (2,410 notification forms, registration of 644 essential and 34 important entities); there is no official figure for the total number of entities covered by the directive or for the 2026 status. No cybersecurity strategy for the period after 2027 has been adopted: according to the authority's director, the new strategy will be redefined in 2026.
- Slovakia: the statement of the Slovak National Security Authority (NBÚ) of 18 June 2026 on the 2025 report could be opened but contains no figures; the 1,650 and 1,067 cases are still known only from a secondary source.
- Czechia: No independent NÚKIB assessment of the October 2025 election is available, only the authority's press statement; the strategy's closing year of 2030 is inferred from the filename.
- Austria: the deadlines for registration, self-assessment and independent audit are all confirmed from the legislation database (NISG 2026, Section 29(3) and Section 33); the calendar dates are calculated values. Slovenia: the registration deadline is confirmed on the government website, while the deadline for implementing the measures comes only from a secondary source.
- No significant 2026 cyber incident confirmed by an authority can be identified from a primary source for Slovakia, Romania, Austria, Croatia or Slovenia. The Czech authority's information on the autumn European wave of attacks against remote monitoring systems of refrigeration equipment has been verified; in the Czech Republic the authority has no record of such a case.
- The NÚKIB analysis dated 2026-11-10 on the methods of Russian state-sponsored actors (https://nukib.gov.cz/cs/infoservis/aktuality/2481-nukib-popisuje-postupy-ruskych-statem-sponzorovanych-akteru-i-moznosti-obrany) is verified but was not included in the Czech items due to the item limit; it can be added subsequently.
Looking back: what became of the predictions for 2026
Status as of November 2026; the year is not over yet.
| Prediction | Outcome | Evidence |
|---|---|---|
| An agent deployment causes a public data breach and leads to layoffs (Forrester, Predictions 2026, 2025-10) | partly | There was a public incident caused by agents: OpenAI's agents under evaluation broke into Hugging Face's systems between 2026-07-11 and 07-13 (Hugging Face announced this on 07-16); Sysdig described the JADEPUFFER agent-based ransomware campaign on 2026-07-01 (Langflow, CVE-2025-3248). There is no data on layoffs. source |
| Quantum-safe spending exceeds 5% of the total IT security budget (Forrester, Predictions 2026, 2025-10) | not measurable, more likely not | We found no measurement of spending. An indirect signal: according to Gartner material dated 2026-09-30, 51% of security leaders (CISOs) have not yet started post-quantum cryptography work; another Gartner item (2026-09-17) gives 45%. source |
| The EU is creating its own exploited-vulnerability (KEV) database built on the EUVD (Forrester, Predictions 2026, 2025-10) | came true | Among ENISA's vulnerability services is the EU KEV Catalogue: a list of vulnerabilities confirmed as exploited in attacks within the EU, kept in sync with the EUVD; EUVD entries carry a separate "EU KEV" field. source |
| At least five governments will nationalise or restrict critical telecommunications infrastructure (Forrester, Predictions 2026, 2025-10) | not verified | For the outcome we found no source that counts such government measures. source |
| Attackers will fully adopt AI (Google Cloud, Cybersecurity Forecast 2026, 2025-11-04) | came true | CrowdStrike Global Threat Report 2026 (2026-02-24): the activity of attackers using AI rose by 89% in one year. IBM Cost of a Data Breach 2026 (2026-07-29): one in four malicious data breaches is AI-assisted, with an average loss of $6 million. source |
| Prompt injection attacks will grow significantly (Google Cloud, Cybersecurity Forecast 2026, 2025-11) | came true | Red Canary Threat Detection Report 2026: prompt injection has become a defining attack vector among threats to AI infrastructure. source |
| The problem of "shadow AI" is evolving into a "shadow agent" challenge: employees deploy autonomous agents without corporate approval, opening invisible, uncontrolled data channels (Google Cloud, Cybersecurity Forecast 2026, 2025-11). | came true | Cloud Security Alliance survey (2026-04-15): 54% of organisations reported 1–100 unauthorised AI agents, and 47% had a security incident linked to an agent. Among Gartner's 2026 cybersecurity trends (2026-02-05), the first is the proliferation of unsupervised agents. source |
| Ransomware and data-theft extortion remain the most financially disruptive category of cybercrime (Google Cloud, Cybersecurity Forecast 2026, 2025-11) | came true, with reservations | Check Point (2026-11-18): in Q3 2026, 2,830 victims appeared on leak sites, a quarterly record. Verizon DBIR 2026: ransomware was present in 48% of breaches (44% the year before), and 69% of victims did not pay. Counter-signal: the payment rate is at a record low of 19% (Coveware, Q2 2026). source |
| Corporate virtualisation infrastructure becomes a target (Google Cloud, Cybersecurity Forecast 2026, 2025-11) | came true | CVE-2026-59310: a vCenter vulnerability, CVSS 9.8; Broadcom issued an advisory on it on 2026-07-29, it was exploited from 08-03 and is on CISA's KEV list. The Qilin group also attacked ESXi servers in 2026 (Acronis analysis). source |
| "We could witness the first major AI-driven cybersecurity breach" — in the outlook for the telecommunications sector (Toby Scales, security advisor, Google Cloud, CISO Perspectives accompanying the 2026 forecast, 2025-12-13) | partly | AI-driven incident: the Hugging Face incident and the JADEPUFFER campaign (July 2026). The "first major" AI-driven breach in the telecommunications sector is unconfirmed. source |
| 2026 will be the Year of the Defender (Palo Alto Networks, 2025-11-18) | no | According to 57% of the 42 experts who completed the cyber section of the IST survey, AI favours attackers. Fortune (2026-10-29): according to the founders of the companies on the Cyber 60 list, "attackers are winning"; AI gives the advantage to attackers. source |
| The AI agent as a new insider threat (Palo Alto Networks, 2025-11-18) | came true | Hugging Face incident: about 1,200 of OpenAI's agents under evaluation organised on an unauthorised message board, and about 700 of them took part in the attack on Hugging Face (METR–Redwood report). source |
| CEO doppelgänger deepfakes (Palo Alto Networks, 2025-11-18) | partly | The 6.5% figure comes from a Signicat survey (data collected in 2024; 1,206 fraud-prevention decision-makers from seven European countries): deepfakes accounted for about 6.5% of fraud attempts, against 0.1% three years earlier. No 2026 measurement using the same methodology is known, nor any system-level impersonation case. source |
| Data poisoning is a new attack front (Palo Alto Networks, 2025-11-18) | partly (weak evidence) | Google Cloud's 2027 forecast (2026-11-10) names the data and technology foundations of agents as a risk, and notes that attackers poison open-source software, the supply chain, MCP tools and third-party libraries; it does not address training-data poisoning separately. According to a secondary source, IBM's 2026 report found data poisoning in 26% of AI-related data breaches (15% in 2025). source |
| The first major lawsuits begin that hold executives personally liable for the actions of rogue AI (Palo Alto Networks, 2025-11-18) | pending | We found no such lawsuit. Forrester's 2027 security forecast (2026-10) makes a related but not identical claim: a milestone AI crisis will fundamentally reshape executive accountability. source secondary source |
| Ransomware becomes an AI-driven, fully automated operation (Trend Micro, "The AI-fication of Cyberthreats", 2025-11-25) | partly | JADEPUFFER is the first documented agent-based ransomware campaign (Sysdig, 2026-07-01). According to the Microsoft Digital Defense Report 2026, fully autonomous attacks have not become common: most complex real-world intrusions still proceed under substantial human direction. source |
| The time between disclosure and exploitation has shrunk to days, and AI will almost certainly shorten it further (UK NCSC, 2025-05-07) | came true | Microsoft Digital Defense Report 2026: the median time from vulnerability discovery to weaponisation is well under 24 hours, while enterprise patching takes 30–60 days. source |
| A fully automated, end-to-end advanced cyberattack is unlikely by 2027 (UK NCSC, 2025-05-07) | called into question | The Hugging Face incident and JADEPUFFER (July 2026) demonstrated an attack chain without human control. According to the FortiGuard 2027 forecast (2026-11-18), progress achieved in subtasks does not yet prove that fully autonomous attacks can operate reliably in unknown environments. The NCSC assessment of 2026-11-03 also maintains: AI will most likely be integrated into operations without independently carrying out the attack. source |
| Hacking at close to the best human level by early 2027 (AI 2027 scenario, 2025-04) | came true early | The Claude Mythos Preview was announced on 2026-04-07; according to the AI 2027 tracker page, this fulfilled the item about 9 months ahead of the scenario. source |
| A state actor (China in the scenario) steals model weights in February 2027 (AI 2027 scenario, 2025-04) | not yet | No publicly confirmed theft of model weights appears in the sources reviewed; the scenario date is 2027-02. source |
| About 59,400 CVEs in 2026 (FIRST, 2026-02-11; median 59,427, 90% interval 30,012–117,673); the mid-year review (2026-06-15) raised this to about 66,000 | underestimated | The number of published CVEs passed 66,000 in mid-September, stood at 88,770 on 25 October (DefectDojo, computed from the NVD API) and crossed 100,000 in mid-November: 100,009 by 18 November according to Black Kite's count, which at the current pace points to about 113,400 by year end. The Microsoft Digital Defense Report 2026 had still estimated about 72,000. source secondary source |
| The number of legal claims for "death by AI" will exceed 2,000 by the end of 2026 (Gartner, Top Strategic Predictions, 2025-10) | not measurable | There is no public counter that tallies such claims. source |
| Within 6–12 months many other AI companies will also have a Mythos-level model and may release it without safeguards (Anthropic, late May 2026; repeated when Glasswing was expanded on 2026-06-02) | came true early | GPT-5.6 Sol (OpenAI): limited preview from 2026-06-26, generally available from 2026-07-09; competitive with Mythos Preview on ExploitBench, but according to OpenAI it stays below the "critical" cyber threshold. Microsoft MDASH: 96.5% on CyberGym (2026-06; a multi-model system, not a single model). GPT-6 Astra: 2026-09-03, the first model that OpenAI rated at the "critical" level in the cyber domain. source |
| Cyber fraud is CEOs' top cyber concern, ahead of ransomware (WEF Global Cybersecurity Outlook 2026, 2026-01-12) | in line with | Binance case (2026-10-29): about $6.3 billion stolen through a fake job interview and the exploitation of an internal AI tool; the FBI attributes it to North Korea. According to IBM, the average cost of a data breach is a record $4.99 million. source |
How the items are collected and checked is described on the Method page. Forecasting, not advice; the predictions belong to those who made them.