Empirical Security Raises $25M for AI-Powered Threat Prediction Platform

Chicago-based cybersecurity startup Empirical has closed a $25 million Series A funding round, bringing total capital raised to $37 million. The round was led by Brightmind Partners, with participation from Costanoa Ventures, Hyde Park Angels, and other investors. Founded in 2024, the company plans to deploy the new capital to expand development of its predictive security platform designed for the agentic AI era.

Empirical's product portfolio centers on two flagship offerings: Foundation, a global cybersecurity model that continuously monitors more than 18,000 exploited CVEs to forecast emerging threats, and Radiant, a predictive engine that filters those signals down to the vulnerabilities most relevant to each customer's specific environment. By layering AI-driven forecasting on top of evidence-based risk analysis, the company claims its models cut through background noise and help security teams prioritize remediation in technology, healthcare, and financial services sectors.

The founding team brings deep credibility to the venture. CEO Ed Bellis and CTO Michael Roytman previously built Kenna Security, where they helped establish the risk-based vulnerability management category. Chief Data Scientist Jay Jacobs co-created the Exploit Prediction Scoring System (EPSS), a widely adopted framework for estimating the likelihood that a CVE will be weaponized. "We helped pioneer the category of risk-based vulnerability management, but it became clear that defending against AI-driven threats and the growing volume of potential exploits would require a fundamentally new approach," Bellis said. Organizations looking to harden their external attack surface alongside these predictive capabilities can run a quick port scanner to identify exposed services, or validate certificate hygiene with an SSL/TLS checker.

For security teams tracking supply-chain and infrastructure risk, Empirical's funding signals continued investor confidence in predictive, data-driven defense models over traditional signature-based detection. Practitioners interested in auditing their own digital footprint can pair enterprise threat intel with a WHOIS lookup to verify domain ownership and registration details that often appear in threat actor reconnaissance.