- London-based Geordie AI has raised $30M in a Series A led by Balderton Capital, valuing the company at approximately $180M post-money — believed to be the largest Series A for a cybersecurity startup in Europe to date
- The startup reported 1,300% ARR growth in the first five months of 2026 as enterprises rush to govern AI agents they have already deployed but cannot adequately monitor
- The cybersecurity agentic AI market stood at $1.83B in 2025 and is forecast to reach $7.84B by 2030, growing at a 33.83% CAGR, according to Mordor Intelligence
Eighty-two percent of enterprises already run AI agents. Only 44% have policies in place to secure them. That gap – between deployment and governance is the business Geordie AI was built to close.
Founded in early 2025 in London by Henry Comfort (CEO), formerly COO Americas at Darktrace through its IPO and $5B exit, Hanah-Marie Darley (Chief AI and Product Officer), former Director of Security and AI Strategy at Darktrace with a decade in US government intelligence, and Benji Weber (CTO), former Senior Director of Engineering at Snyk, Geordie has raised $30M in a Series A led by Balderton Capital.
James Wise, Partner at Balderton, will join the company’s board as part of the investment. Crosspoint Capital joins as a new investor, alongside follow-on funding from existing backers General Catalyst and Ten Eleven Ventures. Total funding now stands at $36.5M, following a $6.5M seed round in September 2025 co-led by General Catalyst and Ten Eleven Ventures. The company currently employs 37 people across London and New York, with headcount expected to reach 50 within three months.
Why enterprises are flying blind on AI agents
Geordie’s platform gives enterprise security teams real-time visibility into which AI agents exist inside their organisation, what data and tools those agents can access, how they behave, and what risks they introduce across enterprise systems. Its runtime remediation suite, Beam, lets security teams constrain agent behaviour without blocking deployment. The distinction from traditional cybersecurity tools is architectural: those tools were built for rule-based software that behaves predictably; AI agents make non-deterministic decisions, evolve over time, and operate across dozens of internal systems simultaneously — none of which existing platforms were designed to handle.
1,300% ARR growth in five months
The company’s growth reflects genuine enterprise urgency. Geordie is now deployed across roughly 30 customer environments, including AlphaSense, where it covers tens of thousands of agents, and Owkin, which runs hundreds of AI agents across more than 50 petabytes of biomedical data. Owkin’s internal risk modelling found that a single Geordie proof-of-concept deployment helped prevent exposure valued at between $12M and $13M. ARR grew 1,300% in the first five months of 2026 alone. In April, Geordie won the RSAC 2026 Innovation Sandbox, cybersecurity’s most closely watched startup showcase.
“Organisations that can safely approve and deploy AI agents are gaining a new competitive advantage,” says Comfort. “Geordie gives enterprises a defence-in-depth approach so they can deploy agent systems safely at scale.”
“AI agents are becoming the operating system of the modern enterprise, but governance and security infrastructure hasn’t kept pace,” says Wise. “Geordie gives security teams the visibility and control they need to deploy agents confidently at scale.” Balderton has previously backed Revolut, Darktrace, and GoCardless — three companies that each became category leaders in their respective verticals.
The competition and what makes Geordie different
In the AI agent security market, Geordie’s closest competitors include Lakera, which raised $20M in a Series A led by Atomico in July 2024 to secure generative AI applications, and Protect AI, which raised $60M in a Series B led by Evolution Equity Partners in August 2024 to expand its ML security platform. HiddenLayer raised $50M in a Series A in September 2023 to protect AI models from adversarial attacks. Where those platforms focus primarily on securing AI models and prompts at the input layer, Geordie focuses on runtime behaviour – monitoring what agents actually do once deployed, not just what instructions they receive.
The harder question the Geordie raise surfaces is not whether enterprises need agent governance – 73% of CISOs say they are critically concerned about AI agent risks, yet only 30% have mature safeguards in place. The question is whether a $180M company can build the category fast enough before the hyperscalers – Microsoft, Google, and AWS – decide that agent governance belongs bundled inside their own platforms.
