Nigerian Founder Launches ADT, a New AI Model for Cyber Defense
Glemad, an AI security and research company, today announced the...
Glemad, an AI security and research company, today announced Autonomous Defence Transformers (ADT) – a new class of AI models pretrained specifically for security reasoning and autonomous defense. This marks the first time frontier-scale AI models have been explicitly optimized for defensive reasoning rather than adapted from general-purpose systems.
“Security has always been reactive. We built ADT to change that – to give organizations the ability to reason about threats continuously and act at machine speed, without sacrificing safety or accountability.”
– David Idris, CEO and Founder, Glemad
The Problem: Human-Paced Defense Fails at AI Speed
Modern attackers operate at machine speed, automate reconnaissance, and adapt behavior dynamically. Yet most security systems remain fundamentally reactive – collecting telemetry after actions occur, triggering detections after damage begins, and gating response through human triage operating under time pressure. The result is a structural mismatch: attackers complete their objectives in the gap between detection and response, while defenders struggle to maintain security invariants over infrastructure state.
The Defense-Native Breakthrough
Since 2020, Glemad has pursued a fundamental thesis: security should be treated as a continuous reasoning-and-control problem, not an alerting-and-documentation problem. This led to ADT – AI systems structurally different from conventional security approaches. ADT models are trained on security-specific corpora from the ground up. Attack telemetry, compliance frameworks, infrastructure semantics, and adversarial tradecraft are embedded as first-class concepts during pretraining – not added later through fine-tuning.
Production Results
PulseADT, the operational implementation of ADT models, is currently protecting over 680,000 assets while processing 1.8 million security events per second.
• Mean Time to Detect (MTTD):0.8 minutes vs. 287 minutes industry average (359x faster)
• Mean Time to Contain (MTTC):3 ...
