Cybersecurity startup Empirical Security, which specialises in building AI-powered defences tailored to enterprise infrastructure, has secured $12m in seed funding.
The investment round was led by Costanoa Ventures, with additional participation from DNX Ventures, Sixty Degree Capital, HPA, and notable strategic investors including Jonathan Cran, Wade Baker, and Gerhard Eschelbeck. Cran is the founder of Intrigue, which was acquired by Google/Mandiant, Baker created the Verizon Data Breach Investigations Report, and Eschelbeck previously served as CTO of Qualys and CISO at Google.
Empirical Security develops artificial intelligence models tailored to the specific systems and risks of each enterprise, offering an alternative to standardised cybersecurity tools. By combining local intelligence with global attack data, the company enables more precise threat prioritisation and risk management.
The newly acquired funds will be used to advance Empirical’s dual-model AI architecture, which builds both global and local cybersecurity models. These local models are trained on an organisation’s unique data and infrastructure, helping security teams identify the most pressing threats with high accuracy.
Empirical’s platform operates on a two-model approach. Global models ingest and learn from approximately two million exploitation events daily to stay ahead of evolving threats. Meanwhile, local models adapt that intelligence to each client’s operations and threat profile. This allows chief information security officers (CISOs) to make evidence-backed decisions and respond more effectively to attacks.
Ed Bellis, co-founder of Kenna Security, has joined Empirical as CEO. He reunites with co-founder and CTO Michael Roytman and co-founder and chief data scientist Jay Jacobs—together, the trio previously led Kenna Security and helped establish the risk-based vulnerability management sector.
Empirical Security CTO Michael Roytman said, “Today’s cyber attacks are custom-built using AI and your own infrastructure against you. Defending with generic, one-size-fits-all models is a start, but only custom, localised models – trained on your data and environment – can close that gap.”
Costanoa Ventures general partner John Cowgill said, “We backed Ed, Michael, and Jay at Kenna, where they pioneered the risk-based vulnerability management movement. With Empirical, they’re doing it again—but replacing generic risk scores with local AI models that tailor scores to each enterprise. We’re thrilled to partner with them on what we believe is a 10x bigger opportunity to transform how security teams prioritise and act on risk.”
Read the daily FinTech news here
Copyright © 2025 FinTech Global
