Empirical Security Launches Foundation V2 to Tackle AI-Driven Vulnerabilities

News related to:Empirical Security · 2 min read

Empirical Security has launched Foundation V2, an advanced vulnerability prediction model designed to help security teams navigate the rapidly evolving landscape of AI-driven vulnerabilities. The new model significantly enhances the company's ability to predict which vulnerabilities are most likely to be exploited, offering a 9.5 times improvement in predicting exploitation within the first seven days of a vulnerability's discovery.

Foundation V2 addresses a critical gap in traditional severity scoring systems, which often fail to keep pace with the speed at which AI uncovers new vulnerabilities. These systems, based on legacy tools, lack the contextual understanding and opinionated approach necessary to prioritize threats effectively. Foundation V2, on the other hand, provides a calibrated probability of exploitation, updated in real-time, enabling security teams to focus on the most dangerous threats.

The new model introduces nearly 300 additional inputs and three new exploitation sources, making it more robust and accurate. According to the release, Foundation V2 outperforms its predecessor in several key areas: - It is 24% better at predicting exploitation in the next 30 days across all Common Vulnerabilities and Exposures (CVEs). - It is 2.2 times better at spotting vulnerabilities that return to exploitation after more than a month of silence. - It is 1.9 times better at recognizing when exploitation attempts are about to halt.

These improvements translate into significant operational efficiency for security teams. With a fixed budget of 7,500 fixes, teams prioritizing with Foundation V2 can address vulnerabilities exploited in the wild within 30 days 93% of the time. This is a marked improvement over the prior version of the model, which managed to fix such vulnerabilities only 60% of the time.

To illustrate the effectiveness of Foundation V2, Empirical Security conducted a comparative analysis of the top 10,000 CVEs as of September 30, 2026. Using the same budget, the company's model outperformed leading vulnerability scanners: - Foundation V2's list held 6,995 vulnerabilities exploited in the wild within 30 days, achieving a 93% hit rate. - Qualys QVS's list held 3,354, or 45%. - Tenable VPR's list held 1,975, or 26%.

In short, for the same effort, Foundation V2 closed twice as many exploited vulnerabilities as Qualys and three and a half times as many as Tenable. This underscores the critical importance of prioritizing based on probability rather than relying solely on severity scores, which often fail to reflect the true risk of a vulnerability.

Ed Bellis, co-founder and CEO of Empirical Security, emphasized the transformative impact of AI on exposure management.

Foundation V2 is now available to all Empirical Security customers, marking a significant step forward in the company's mission to help organizations stay ahead of the rapidly evolving threat landscape.

Start filing today

One press release free every week. No card required.

Create a free account