Trent AI, a startup focused on securing agentic AI systems, has raised $13 million in Seed funding led by LocalGlobe and Cambridge Innovation Capital.
The round also included participation from senior figures at OpenAI, Spotify, Databricks, and Amazon Web Services. The funding supports the development of a new security layer designed to address risks associated with autonomous AI agents.
What The Company Does
Founded in 2025 by Eno Thereska, Neil Lawrence, and Zhenwen Dai, Trent AI is developing a platform that secures AI agents through a multi-layered, context-driven approach.
The system is built around a proprietary judgement layer combined with reinforcement learning, enabling specialised security agents to continuously monitor AI systems. These agents operate across the lifecycle of AI deployments, identifying vulnerabilities, managing risks, and adapting to evolving threats.
Unlike traditional security tools, which are often static, Trent AI’s platform is designed to improve over time through feedback loops, allowing it to refine threat detection and response as systems scale and become more complex.
Advertisement
Market Context / Industry Background
The rise of agentic AI, where autonomous systems interact and make decisions independently, is introducing new security challenges. As organizations increasingly deploy multi-agent workflows, traditional governance frameworks are struggling to keep pace.
Industry data indicates that a large proportion of companies are planning to adopt agent-based AI systems in the near term, yet many lack dedicated security infrastructure tailored to these environments. This gap creates potential vulnerabilities, particularly in systems where multiple agents operate simultaneously and influence each other’s outputs.
Trent AI is entering a developing segment of the cybersecurity market focused specifically on AI-native threats, where continuous monitoring and adaptive protection are becoming critical requirements. As this segment evolves, organizations are expected to require more specialized tools capable of addressing dynamic interactions between agents, ensuring visibility, traceability, and control across increasingly complex and autonomous digital ecosystems. This growing demand underscores the need for robust, scalable frameworks that can adapt to rapidly changing threat landscapes, while maintaining operational efficiency and minimizing systemic risks.
Founder / Investor Commentary
Chief Executive Officer Eno Thereska highlighted the mismatch between rapid AI adoption and slower security adaptation, noting that many development teams lack frameworks designed for agent-based systems. He emphasized that securing these environments requires new approaches capable of operating alongside autonomous workflows.
From the investor perspective, Saul Klein pointed to the increasing complexity of AI-driven systems and the corresponding rise in security risks. He noted that building long-term security foundations for agentic AI is becoming essential, and positioned Trent AI as a company combining academic expertise with practical system-building experience.
The investor group also includes individuals such as Joaquin Quiñonero Candela, Avinash Bhat, Ippokratis Pandis, and Tony Jebara, reflecting a mix of technical and operational backgrounds.
Growth Plans / Use Of Funds
The funding will be used to further develop Trent AI’s platform, expand its engineering team, and support early customer deployments. The company is working with initial users including Canopy, Commscentre, ML@Cam, Qbeast, and Weblogic.
These early adopters report improvements in system visibility, faster identification of vulnerabilities, and clearer remediation processes, indicating initial validation of the platform’s approach.
As adoption of agentic AI accelerates, Trent AI aims to position itself as a foundational security layer embedded within AI development and deployment workflows.
About Trent AI
Trent AI is a security startup focused on protecting agentic AI systems. Founded in 2025, the company develops a platform that uses reinforcement learning and a proprietary judgement layer to power specialised security agents. Its mission is to provide continuous, adaptive protection for AI systems, supported by a team with experience across cloud infrastructure, large-scale platforms, and academic research.
