Corma has raised $60 million in seed funding to build a foundation model specifically for defensive cybersecurity, targeting a widening gap between rapidly advancing AI-powered attacks and the ability of existing security systems to detect and stop them. Sequoia Capital led the financing, with participation from Khosla Ventures and Coatue. Founded in 2025, Corma describes itself as a frontier AI lab focused exclusively on defensive cybersecurity. The company operates from Tel Aviv and San Francisco and is already working with Fortune 100 and Fortune 500 organizations. Corma’s central thesis is that advances in general-purpose AI have disproportionately benefited attackers. Frontier models have become increasingly capable at coding, software reasoning, vulnerability research and multi-step tool use, capabilities that can also be applied to identifying vulnerabilities and carrying out cyberattacks. But Corma argues that defending large enterprise environments presents a fundamentally different AI challenge. Defensive systems must analyze enormous quantities of audit logs, security events, network flows, and other information, correlate weak signals over extended periods, and maintain consistent reasoning across thousands of decisions. Corma conducted hundreds of simulations using realistic enterprise environments modeled after Fortune 500 organizations and incorporating the dozens of security tools commonly deployed by large companies. Leading AI models, including systems from OpenAI and Anthropic, were first instructed to act as attackers and establish persistent threats inside the simulated enterprise environments. The same models were subsequently asked to operate as defenders and identify and remediate the threats they had created. According to Corma’s research, AI attackers succeeded in 88% of the simulations, while AI defenders detected only 12% of the threats. The company said that in nearly every case, the same model capable of executing an end-to-end attack could not successfully defend against that attack when operating from the opposite side. Corma believes that imbalance demonstrates why adapting
Corma Raises $60 Million Seed To Build Defensive <b>Cybersecurity</b> Foundation Model As AI ...
Read the original article
pulse2.com →