[ad_1]
London-based Trent AI has raised $13 mn, in seed funding as it comes out of stealth with a layered security product built for the agentic AI market. LocalGlobe and Cambridge Innovation Capital led the round.
Trent AI enters the market as companies push AI agents and autonomous workflows into production faster than security teams are able to adapt.
Investors also included OpenAI technical staff member Joaquin Quinonero Candela, former Stripe data infrastructure head and current AWS director Avinash Bhat, Databricks distinguished engineer Ippokratis Pandis, and former Spotify vice president of engineering and AI leader Tony Jebara.
According to the company, most development teams using these systems still lack a security framework built for them.
Chief executive officer and co-founder Eno Thereska said organisations are deploying AI agents and autonomous workflows at a pace security teams are struggling to match.
He said development teams using these systems often operate without a security framework suited to how agentic infrastructure works.
Organisations are deploying AI agents and autonomous workflows faster than their security can adapt, and most development teams using these agents and workflows have no security framework designed for their systems
Eno Thereska, co-founder and CEO of Trent AI
Cambridge Innovation Capital partner Ian Lane said adoption of autonomous agents is moving faster than enterprise security readiness.
As these systems start making decisions across critical infrastructure, he said companies need a new layer of tools to govern, monitor, and enforce safe behaviour. In his view, Trent AI is in a strong position to define that category.
Founded in 2025, Trent AI says it aims to reshape agentic AI security through a context-driven model. Its proprietary judgement layer and reinforcement learning technology sit behind a set of specialised security agents.
By coordinating those agents across customer workflows, the company wants security to become a continuous part of agent development rather than a check added later.
The founding team brings a heavy technical profile. Trent AI was launched by Eno Thereska, formerly a distinguished engineer at Alcion, AWS, and Confluent, Neil Lawrence, DeepMind professor of machine learning at the University of Cambridge and former Amazon machine learning director, and Zhenwen Dai, previously a machine learning scientist at AWS and a senior research manager at Spotify.
Saul Klein, co-founder and executive chairman of Phoenix Court, home of LocalGlobe, said the rise of agents is bringing a fresh set of security threats with it. He said the market has reached the point where long-term security foundations for agentic systems need to be built now, and argued Trent AI is well placed to do that because of its mix of academic depth and practical experience.
The timing lines up with broader market demand. Deloitte’s 2026 State of AI report found that almost three in four companies, or 74%, plan to deploy agentic AI within two years.
Yet only 21% said they already have a mature governance model for autonomous agents. That gap gets riskier in environments where multiple agents interact, since a weak point in one system can spill into the wider stack.
Trent AI says its platform secures agents across the full lifecycle. Each cycle is designed to make its own security agents better at understanding the systems they protect.
As the feedback loop tightens, the company says judgement improves and mitigation becomes more accurate, giving development and security teams a quicker path to safer deployment.
Early-access users including Canopy, Commscentre, ML@Cam, Qbeast, and Weblogic have reported better visibility into their security posture, along with audit reporting, fast identification of vulnerabilities, cleaner remediation scope, and adaptive feedback.
Avinash Bhat, director at AWS, said agentic systems are quickly becoming part of the software stack even though the security infrastructure around them is still early. He said Trent is building the kind of foundation teams will need to run these systems safely at scale.
[ad_2]
Click Here For The Original Source.
