Vulnpocalypse

Cyber defence vendors are increasingly seeing threat actors use Large Language Models (LLMs) to work out how to weaponise known vulnerabilities that haven’t yet been exploited in the wild.

This is a scenario more security teams should be thinking about going into 2026/2027.

Traditionally, security teams prioritise patching based on active exploitation, so if a vulnerability is known but hasn’t been used by attackers, it often falls lower on the urgency scale.

But AI could change that equation fast.

If attackers can systematically use LLMs to uncover reliable exploit paths for previously unexploited vulnerabilities, defenders could face a wave of “new” zero-day style attacks at unprecedented scale, the nightmare scenario being referred to as a VULNPOCALYPSE.

This shift is why AI-driven defence is becoming increasingly important. Traditional security tools often rely on known threat signatures and attack databases, which can leave organisations exposed to new and evolving threats. AI-driven security takes a different approach, identifying unusual behaviour and suspicious activity that may indicate an attack before it has been formally classified.

The future of cyber defence may depend less on what we already know and more on how quickly we can detect what shouldn’t be happening.

To help mitigate the Zero Day Threat, NG-IT can help organisations to deploy AI-based solutions that identify threats, reduce exposure, and accelerate remediation.