Artificial intelligenceJune 19, 2026· via AI News

Cumulo: The U.K.’s first sovereign, AI-driven SOC platform for zero-day threats

Cumulo: The U.K.’s first sovereign, AI-driven SOC platform for zero-day threats

The U.K. has just gained a new weapon against fast-moving cyber threats. e2e-assure, a SOC-as-a-service provider based in Abingdon, has launched Cumulo, a platform described as the country’s only sovereign, AI-first solution designed to secure both IT and operational technology environments. Unlike traditional security operations centres that react to alerts long after an attack begins, Cumulo is built to identify vulnerabilities and emerging threats before incidents escalate.

A platform built for speed and sovereignty

Cumulo integrates digital twin technology with customer-dedicated AI models, enabling continuous context building as data is generated. The platform operates alongside a SIEM that maintains a deterministic record of every event, while AI runs in parallel to enhance detection and response capabilities. This design allows Cumulo to apply new threat intelligence immediately as detection rules, effectively creating a “zero-day SOC” that can neutralise emerging threats faster than conventional systems.

The platform is developed and hosted in the U.K., aligning with calls from GCHQ for a national cyber defence capability that leverages agentic AI at machine speed. Critically, Cumulo avoids full AI autonomy by keeping SC-cleared security teams at the centre of every decision. Human oversight remains integral, ensuring that while AI accelerates detection, final judgements are made by experts.

Digital twins and localised AI for safer defence

Cumulo creates a continuously updated digital twin of each customer environment through passive discovery across IT and OT systems. This allows organisations to simulate attacks, identify risks before exploitation, and preserve analytical integrity—especially valuable in critical infrastructure, where live testing is often impractical. Each customer receives a dedicated local large language model, trained on their specific environment to provide context-aware reasoning without relying on external cloud AI services. This approach not only reduces latency but also ensures sensitive security data remains under the organisation’s control, reinforcing sovereignty and compliance.


Source: AI News. AI-assisted editorial synthesis — TechnoExpress.

Read the original source on AI News →

← Back to home