Germany’s Federal Office for Information Security has raised its assessment of the country’s cyber threat environment from “tense” to “tense to critical”, warning that artificial intelligence is helping attackers find and exploit weaknesses at a faster pace. The BSI’s 2026 annual report, published on November 19, covers the period from July 1, 2025, to June 30, 2026. It identifies AI-assisted vulnerability discovery, growing attack surfaces and weak basic security practices as factors behind the deterioration. The assessment is a national threat review, not a claim that every organisation has been breached.
The agency introduced the BSI Cyber Index in the report as a way to combine several dimensions of the threat environment into a single measure. The index averaged 170 points during the reporting period, 50% higher than in the previous period, according to the BSI’s summary. Reuters separately reported a mid-year reading of 238 points, compared with 124 at the same point in 2025. The average and the mid-year figure describe different points in the index’s development and should not be treated as interchangeable measures.
The time between disclosure and exploitation is shrinking
The BSI said an average of 160 new IT vulnerabilities became known each day during the reporting period, 35% more than a year earlier. More concerning for defenders was the decline in the time between public disclosure of a vulnerability and the first known exploitation. The agency reported an average of 53 days in the previous reporting period, 25 days in the latest period and eight days in the second quarter of 2026. That final figure indicates that the gap between a security warning and exploitation can become extremely short for some vulnerabilities.
The change affects the way organisations plan patching and vulnerability response. A company that waits for a monthly maintenance window may leave a known weakness exposed for weeks, while attackers may already be scanning for systems that have not received a fix. The BSI’s figures do not mean every newly disclosed flaw is exploited within eight days. They describe the reported time-to-exploit trend in the agency’s monitoring. The practical implication is that organisations need a process for identifying exposed assets, prioritising high-risk vulnerabilities and deploying emergency updates outside routine schedules.
Artificial intelligence is a factor in that acceleration because modern tools can help identify vulnerable software, test possible inputs and generate code that uses a weakness. The BSI warned that advanced AI systems could allow attacks at quantities and speeds that were previously difficult to achieve. Its president, Claudia Plattner, said the technology had changed the scale and complexity of cyber aggression. The report also anticipates more activity from autonomous AI agents, although the public summary does not claim that all current attacks are autonomous or that AI is the sole cause of the increase.
Small businesses and local authorities remain exposed
The BSI said small and medium-sized businesses and local authorities remain particularly vulnerable targets. Those organisations often operate a mixture of modern cloud services, legacy applications, connected devices and systems supplied by external vendors. They may not have the specialist staff or budget available to large infrastructure operators, and security updates can compete with other operational priorities. The report said fewer than half of users of digital applications employed basic security measures, an indication that the problem extends beyond advanced threat techniques to routine account and device protection.
The agency also warned that Germany’s attack surface is expanding as digitisation and connected devices spread across the economy. Each internet-connected device, externally accessible service or supplier integration can create another route into an organisation. The risk is not limited to computers used by employees. Industrial systems, building controls, medical equipment, logistics platforms and consumer devices can be connected to networks that were not originally designed for exposure to internet-scale threats. The BSI’s assessment links this growth in connected infrastructure with a need to improve preventive measures at the same pace.
The BSI says defensive investment is not keeping pace
The report says resilience has improved in some areas, but preventive measures are not being implemented quickly enough to match the expanding threat environment. That gap is important for organisations that believe they are secure because they have purchased security software. Tools cannot compensate for missing asset inventories, unpatched systems, weak identity controls or unclear responsibility for responding to alerts. The BSI’s assessment places the emphasis on operational readiness: organisations need to know what they operate, where it is exposed and who can make decisions when a vulnerability becomes urgent.
For technology vendors, the shorter exploitation window increases pressure to release fixes quickly, communicate affected versions precisely and support customers that cannot upgrade immediately. For operators, it increases the value of automated asset discovery, vulnerability scanning, endpoint monitoring and tested incident procedures. AI can also support defensive work by helping teams analyse code and prioritise alerts, but those systems need access to reliable information and a clear route from detection to remediation. An alert that is not assigned to an owner or acted on promptly does not reduce exposure.
The BSI’s assessment is a warning about the direction and speed of the threat, not a prediction that every German organisation will suffer an incident. The report covers a defined twelve-month period and combines changes in threats, attack surfaces, attacks, impacts and resilience. Its clearest operational message is that the time available to respond to a newly disclosed vulnerability is shrinking. Organisations that continue to rely on slow patch cycles and incomplete inventories risk falling behind attackers that can automate more of the discovery and exploitation process.