Three months after introducing its product, AI cybersecurity startup Corma has raised a $125 million Series A round at a $1 billion valuation, led by NEA, according to the company. Corma offers AI models and agents trained for cybersecurity defense, and it has now raised $185 million in total from backers including Sequoia Capital, Khosla Ventures, and Coatue. Those are serious names, and that kind of money arriving this fast deserves more scrutiny than applause.

The company's timing was not accidental. Its product drew attention after a run of security incidents sent chief information security officers scrambling on October 21, when a critical flaw was found in PHP, a popular web programming language. Another flaw was found in Node.js, the open source JavaScript runtime, and LiteLLM, a proxy for routing requests to multiple LLM providers, disclosed a security incident. The report also notes that other incidents were disclosed that day. The lesson for buyers is that fear sells well when the threats are real, and the report describes the day as showing the kind of onslaught on security that AI has enabled.

That framing is useful, but it also makes a valuation easy to justify. A $1 billion price tag three months after launch is a claim about future results, not about delivered protection. The report does not offer customer counts, revenue, or independent evidence that Corma's models stopped a single attack. Corma says it stands out by being a frontier AI research lab focused on protection, yet a self-description is not a track record. Investors may be right, but the public record here is thin.

The competitive picture is crowded as well. The report points to Kevin Mandia's new startup Armadin, to Anthropic, OpenAI, Microsoft, and other established players, and to what it calls too many other startups to count. The report itself calls this a young market and a necessary one, and it is right that defense needs better tools. But a market where everyone claims to use AI for protection is also a market where differentiation is hard to verify and marketing can outrun results.

What should happen next is straightforward. Security leaders evaluating Corma or any similar vendor should ask for measurable outcomes, including detection rates, false-positive rates, and case studies from named customers willing to be cited. Reporters and analysts should press for the same. A billion-dollar valuation is a signal of investor confidence, but it is not proof that the product works. Until the evidence catches up with the money, the most responsible reaction is to be impressed by the funding and skeptical of the promise.