Trending: On-device modelsSearch
iHeartGeek
iTECH

Corma raises $125 million for defensive cybersecurity AI

The cybersecurity startup says its Series A, led by NEA, lifts total funding to $185 million at a reported $1 billion valuation.

Corma's luminous ring artwork from its official website

Cybersecurity startup Corma says it has raised $125 million in a Series A round led by NEA, taking its total funding to $185 million and putting the company at a reported $1 billion valuation.1

The company announced the round three months after emerging from stealth with $60 million. Sequoia Capital, Khosla Ventures and Coatue are continuing to back the business, according to Corma’s announcement.

What Corma is building

Corma says it is developing foundation models and an AI workforce for defensive cybersecurity. The company’s pitch is that security teams can delegate complex work to software agents that operate alongside them and use their existing security tools.

According to the company, that workforce was already being used by security teams at some large organisations during the wave of vulnerabilities and breaches it calls “Black October”. Corma says demand has been strong enough to create a compute-capacity bottleneck, with more work needed to expand access for customers on its waitlist.1

Those claims come from Corma itself. The announcement does not provide independent performance benchmarks, customer names or a detailed breakdown of how the company’s models compare with established security products. The funding round is the verifiable development; the claims about what the AI workforce can accomplish remain company-reported.

Where the money is going

Corma says the new funding will be used to scale its foundation models, increase compute capacity and selectively expand its team. The company describes the round as a step towards its broader ambition of building “superintelligence for defensive cybersecurity”, but it also acknowledges that much of the work is still ahead.

The timing reflects a wider rush to apply generative AI to security operations. Defenders face a growing volume of alerts, vulnerabilities and routine investigation work, while attackers are also gaining access to increasingly capable automation. A system that can safely handle parts of that workload could be valuable, but the difficult questions are operational: what it is allowed to change, how its actions are reviewed and how customers recover when an automated decision is wrong.

Corma’s announcement does not answer those questions. It says the workforce operates with customers’ existing security tools, but it does not set out the approval controls, audit trail or safeguards governing those actions.

A billion-dollar test

The reported valuation gives Corma a substantial head start in a crowded market. It also creates pressure to show that the company’s foundation models deliver more than a compelling demonstration. Security buyers will need evidence that the system improves response times without introducing new weaknesses, leaking sensitive data or making the work of human defenders harder to audit.

For now, the $125 million round supplies Corma with the resources to pursue that test. The company says new generations of its models are coming soon, while customers waiting for access may still have to wait as compute capacity expands.

Our opinion

Corma has raised enough money to make its defensive-AI thesis impossible to ignore, but the announcement is stronger on ambition than evidence. The funding, backers and valuation are clear; the claims about foundation-model capability are not independently demonstrated here. That distinction matters in cybersecurity, where a confident system that cannot explain its actions is a liability rather than a shortcut. Corma’s next milestone should be transparent results from real deployments, not another grand description of what its workforce might eventually do.