- AI company Prevalent has raised $22 million from Integrity Growth Partners, its first external investment in nine years.
- According to Paul Stokes, he did not seek external investment until the market was ready.
- The team of 200 people is currently extending its knowledge graph technology from cybersecurity to address financial crime.
Paul Stokes and Arun Raj set up Prevalent AI in 2017 after they had sold their earlier cybersecurity company, and Stokes refused any external financing for nine years.
This month, the UK-based cybersecurity startup secured a $22 million investment from Integrity Growth Partners, a growth equity firm headquartered in Los Angeles. This investment marks the company’s first major funding round.
In 2021, Istari, part of Singapore’s Temasek, purchased a minority stake, but this was a secondary transaction and did not raise new capital.
Stokes, co-founder and CEO of Prevalent AI, tells Tech Funding News, “We believed it was possible that, if we raised money too early, we might have been raising it at the wrong time, since the market had not yet recognised what was required.”
Nine years of saying no
Stokes, who is from New Zealand, has for over 20 years worked in the UK in the field of data science for security purposes, initially with the country’s intelligence and defence agencies and later in commercial cybersecurity.
He founded Prevalent AI to cater to large enterprise customers rather than adhere to investor funding cycles.
“We have had a number of large enterprises as loyal customers right from the start, and the product was developed together with them. Because we have concentrated on product development and have devoted all our resources to it instead of raising funds, we have been able to create a solution that meets the needs of major enterprise clients,” he says.
The original group of co-founders had connections with GCHQ, the UK’s signals intelligence agency. Sir Iain Lobban, who was a board member, headed GCHQ from 2008 until 2014. Andrew France, who is also one of the founders, resigned from his post at GCHQ’s Cyber Defence Operations to become CEO of Darktrace, the AI cybersecurity company that was taken private by Thoma Bravo for $5.3 billion in 2024, after first going public on the London Stock Exchange in 2021.
Stokes points out that Prevalent AI’s main customers include banks, telecom companies, and insurers, with over 5,000 employees, and sometimes over 100,000, spread across different regions.
What the knowledge graph does
The product offered by Prevalent AI gathers data from hundreds of its internal tools, security platforms, cloud logs, and identity systems, and combines it into what the company calls a sovereign knowledge graph.
This map of the organisation’s data is continuously updated and deduplicated and is built and hosted under the customer’s control, rather than being located on a shared third-party cloud. Both human analysts and AI agents can make use of the graph to locate information.
Stokes says, “We present it in graphical form so as to make it very contextual, and then let them go on to solve the problem they want to solve.”
The graph was not constructed using large language models, which is a significant technical distinction. “You can’t use AI for that because it would lead to assumptions and deviations in the graph that aren’t accurate. We construct the graph in the same way each time,” he adds.
Even though other providers of AI-native data tools assert that it can be easily copied, nine years’ work on deterministic data reconciliation cannot be readily duplicated by competitors.
Prevalent AI is in a position similar to, rather than in direct competition with, general-purpose graph database vendors such as Neo4j and TigerGraph, since customers typically arrange these for their own security requirements. Instead of having to set up the infrastructure themselves, customers get the reconciliation layer as a managed product designed for complex enterprise security data.
Why now, and what changes
Revenue has doubled each year, but Stokes admits it will be harder to keep up this growth as the company gets bigger. He credits the recent growth not to new sales tactics, but to enterprises realizing that foundation models are “already plenty smart enough” and that the main problem is fragmented internal data.
Gartner says that enterprises will spend $240 billion on information security in 2026 alone and that more than 40% of agentic AI projects will be cancelled by the end of 2027 due to cost, unclear value, and poor risk controls.
As Stokes explained, security teams have to make decisions involving thousands of systems, controls, identities, and data sources that were never designed to work together, and the investment from Integrity Growth Partners will allow them to expand our business from cybersecurity into all the areas of the enterprise where decisions are based on connected and trusted data.
Ryan Anderson, who is the managing partner and co-founder of Integrity Growth Partners, pointed out the founders’ careful handling of capital: “Paul, Arun and the team have created something rare — genuinely differentiated, AI-native technology which the most sophisticated enterprises around the world depend on — and at the same time have shown remarkable capital discipline. Since enterprise security stacks are becoming more complex and since agentic AI is increasing the importance of data quality and depth, the demand for Prevalent AI’s integrated, context-driven foundation will only become greater.”
Nine years of turning down capital gave Stokes leverage most founders never get to negotiate from. It has not yet become clear whether this discipline will persist when the company sets up its first sales team, opens a US office, and competes for trusted enterprise context.
