Megan Harper
In December 2025, the AI research lab Glemad secured a $500,000 private investment to expand the computing infrastructure for its proprietary “Autonomous Defense Transformer” (ADT) model family.
Led by founder and CEO David Idris, Glemad is pursuing a contrarian thesis in the age of generative AI: that critical digital infrastructure requires “security-native” intelligence designed from first principles, rather than general-purpose cloud models wrapped in simple prompts.
General AI models are optimized for broad instruction following, which often fails in high-stakes environments due to latency and hallucination risks. Under Idris’s leadership, Glemad developed ADT as a “digital immune system.” Unlike off-the-shelf APIs, ADT is trained specifically to reason about system configurations, threat patterns, and compliance frameworks in real-time. Through its PulseADT platform, Glemad offers regulated sectors—such as Banking, Healthcare, and Government—a way to automate incident response with “audit-ready” traceability, enabling machine-speed defense that remains transparent and accountable.
The new capital will be used to expand GPU capacity to support larger training runs for ADT-4 and its successors, advancing Glemad’s mission to provide sovereign, high-assurance intelligence for the world’s most sensitive operations.
Market Signal
- The Idris Vision for Sovereign AI: David Idris has positioned Glemad at the forefront of the “Sovereign AI” movement. By building homegrown models that prioritize data sovereignty and local compliance, Glemad is creating a template for how organizations can adopt AI without surrendering control to foreign-hosted black-box systems.
- Defense-First Pretraining: Glemad challenges the trend of fine-tuning generic LLMs. Their approach of “Defense-first pretraining”—using datasets built around infrastructure behavior and threat intelligence—signals a move toward specialized “Foundation Models for Security” that prioritize accuracy over broad creative capabilities.
- From Automation to Autonomy: Idris explicitly distinguishes “Autonomous Defense” (model-driven reasoning) from simple “Automation” (scripted actions). This defines a mature category where the AI doesn’t just execute a playbook, but understands the implications of a threat and the regulatory requirements before taking action.
Key Details
- Company: Glemad
- Funding: $500K (Private Investment)
- Headquarters: Lagos, Nigeria / Sheridan, WY
- Founder & CEO: David Idris
- Core Product: Autonomous Defense Transformer (ADT)
- Use Case: Security-native reasoning and autonomous incident response
