Key Takeaways
- • Corma raises $60 million in Seed funding led by Sequoia Capital, with participation from Khosla Ventures and Coatue, positioning itself as the first frontier AI lab dedicated exclusively to defensive cybersecurity.
- • The company has developed a specialized foundation model trained from the ground up for cyber defense, contrasting with general-purpose AI models that excel at offensive attacks but lack defensive capabilities.
- • Corma deploys AI agents that operate across existing enterprise security tools and processes rather than replacing traditional security software, functioning as "virtual human resources" to augment security teams at Fortune 100 and Fortune 500 companies.
- • Founded in 2025 with headquarters in Tel Aviv and San Francisco, the 20-person startup addresses a critical market gap: as AI-powered cyberattacks accelerate in volume and complexity, enterprise security teams increasingly cannot manage threats manually.
- • The company's business model avoids software replacement—instead providing AI-powered services that work alongside existing security personnel, reducing operational friction while improving security team effectiveness and capacity.
Summary
Corma, a newly emerged startup founded in 2025, has secured $60 million in Seed funding led by Sequoia Capital to build foundation models specifically designed for defensive cybersecurity. The round includes participation from Khosla Ventures and Coatue, according to Fortune's exclusive reporting. Headquartered in Tel Aviv and San Francisco with a team of 20 people in Tel Aviv, Corma is positioning itself as the first frontier AI laboratory dedicated entirely to defensive cyber operations.
The company addresses a fundamental asymmetry in artificial intelligence: general-purpose large language models have become increasingly effective at orchestrating cyberattacks, yet they lack specialized training for defensive operations. Rather than deploying conventional cybersecurity software, Corma has built its own foundation model from scratch, trained specifically on defensive cybersecurity tasks. CEO and co-founder Alon Pluda explained the strategic approach: "We train language models for defensive cyber. The models thus become much better at cyber than any other model."
Corma's deployment model differs significantly from traditional cybersecurity vendors. The company operates AI agents across an organization's existing security tools, processes, and personnel, functioning as supplementary "virtual human resources" rather than replacing existing infrastructure or staff. Pluda emphasized this collaborative positioning: "We don't replace anyone and we are not a product; we provide services. Companies still have thousands of security people and pay a lot for services. I am not replacing them but making them much more successful and much stronger." This approach reduces organizational friction while addressing the critical bottleneck facing enterprise security teams: the sheer volume and complexity of threats now exceed manual handling capacity.
The timing reflects market urgency. As AI-enabled attacks grow faster and more autonomous, enterprises face mounting difficulty managing threats with human-only or conventional tool-based approaches. Corma's AI agents operate end-to-end across security operations, handling repetitive tasks and amplifying the effectiveness of existing security teams. The company is already engaged with Fortune 100 and Fortune 500 enterprises, validating demand for specialized defensive AI capabilities at the enterprise level.
The company addresses a fundamental asymmetry in artificial intelligence: general-purpose large language models have become increasingly effective at orchestrating cyberattacks, yet they lack specialized training for defensive operations. Rather than deploying conventional cybersecurity software, Corma has built its own foundation model from scratch, trained specifically on defensive cybersecurity tasks. CEO and co-founder Alon Pluda explained the strategic approach: "We train language models for defensive cyber. The models thus become much better at cyber than any other model."
Corma's deployment model differs significantly from traditional cybersecurity vendors. The company operates AI agents across an organization's existing security tools, processes, and personnel, functioning as supplementary "virtual human resources" rather than replacing existing infrastructure or staff. Pluda emphasized this collaborative positioning: "We don't replace anyone and we are not a product; we provide services. Companies still have thousands of security people and pay a lot for services. I am not replacing them but making them much more successful and much stronger." This approach reduces organizational friction while addressing the critical bottleneck facing enterprise security teams: the sheer volume and complexity of threats now exceed manual handling capacity.
The timing reflects market urgency. As AI-enabled attacks grow faster and more autonomous, enterprises face mounting difficulty managing threats with human-only or conventional tool-based approaches. Corma's AI agents operate end-to-end across security operations, handling repetitive tasks and amplifying the effectiveness of existing security teams. The company is already engaged with Fortune 100 and Fortune 500 enterprises, validating demand for specialized defensive AI capabilities at the enterprise level.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.