Corma secures $125 million Series A at $1 billion valuation

Corma reaches a $1 billion valuation
AI cybersecurity startup Corma has raised a $125 million Series A round at a $1 billion valuation, three months after it introduced its product and emerged from stealth in August. NEA led the financing, Corma said.
The company has now raised $185 million in total. Its backers include Sequoia Capital, Khosla Ventures and Coatue, alongside the new Series A lead investor.
Corma develops AI models and agents trained for cybersecurity defense. It describes itself as a frontier AI research lab focused on protection, positioning that emphasis as a differentiator in a rapidly expanding field of AI security products.
A crowded day for security teams
Interest in Corma’s offering increased after a series of security events on October 21. A critical flaw was found in PHP, the widely used web programming language, while another flaw was identified in Node.js, the open-source JavaScript runtime environment.
LiteLLM, an open-source proxy used to route requests among multiple large language model providers, also disclosed a security incident that day. Other incidents were disclosed as well, creating a concentrated workload for chief information security officers and their teams.
The cluster of events illustrated the volume of security pressure that AI has enabled. Corma is one of many companies seeking to use AI in defense rather than attack, in a market that remains young but is increasingly necessary for organizations managing an expanding stream of vulnerabilities and incident reports.
AI defense market takes shape
Corma is not alone in building AI-driven cyber defense capabilities. Kevin Mandia’s startup Armadin is another example, while established companies including Anthropic, OpenAI and Microsoft are also active in the area. The field also contains a large number of startups.
The financing gives Corma substantial backing soon after its public debut, but the company enters a competitive market with both specialist newcomers and major AI providers. Its stated focus on frontier research for protection will be central to how it distinguishes its models and agents.
For businesses, the immediate implication is practical: vulnerability and incident workflows need to be able to absorb simultaneous disclosures across widely deployed software, while teams evaluate where AI-based defensive tools could support their existing security operations.

