Skip to content
Business

Corma Hits $1 Billion Valuation With $125 Million Round, Three Months Out of Stealth

NEA led the Series A for the start-up of about 20 people, which trains its own defence models and rates them on a benchmark it built itself.

·2 min read

Prefer Nakoda News on Google

Corma, which builds AI models for cyber defence, raised a $125 million Series A led by NEA at a $1 billion valuation.
Corma, which builds AI models for cyber defence, raised a $125 million Series A led by NEA at a $1 billion valuation.Photo: TechCrunch

Corma, a cybersecurity start-up with offices in Tel Aviv and San Francisco that trains its own AI models to defend corporate networks, has raised $125 million in a Series A round at a $1 billion valuation, TechCrunch reported on Monday. The round comes three months after the company emerged from stealth and brings its total funding to $185 million.

NEA led the round. Sequoia Capital, Khosla Ventures and Coatue, which backed Corma's $60 million seed round in August, invested again, bringing total funding to $185 million. Sequoia led the seed.

The round puts a $1 billion price on a company with about 20 employees and a product that has been on the market for three months. Corma presents itself as a frontier research lab focused only on defence and builds its models from scratch rather than layering a product over systems from OpenAI, Anthropic or DeepSeek. Investors are betting that attackers armed with AI will push large companies to buy defensive tools that work at the same speed.

Chief executive Alon Pluda founded Corma last year, according to Globes. He is a graduate of the Israeli military intelligence's Psagot programme and served in technical roles in Unit 8200 until 2022. He previously co-founded Miggo Security and left after a dispute with his co-founders, the Israeli business daily reported. Corma has offices in Tel Aviv and San Francisco.

The company's models take in telemetry from firewalls, network monitoring software and other security tools to find breaches, then write remediation guidance and can run investigations on their own, alerting staff when a human decision is needed, SiliconANGLE reported. Security teams can steer the system through a Slack chatbot. Corma says its models speed up some breach-response workflows by a factor of 100, and that one of them outperformed GPT-6 Astra and Claude Opus 5.5 on DefenseBench V5, an evaluation Corma built itself. No independent benchmark results have been published.

Pluda said "demand from the world's largest organizations has been overwhelming" and that securing computing capacity fast enough has been one of the company's hardest problems. Corma has increased its server capacity 24-fold since August and plans to spend the new money on more infrastructure and on improving its models.

Globes reported that Corma's annual recurring revenue is estimated in the tens of millions of dollars and that its customers include Fortune 100 and Fortune 500 companies in healthcare, financial services, energy, retail and telecommunications. Corma has not named any of them.

Demand picked up after October 21, when security chiefs scrambled to respond to a critical flaw in the PHP programming language, another in the Node.js runtime and a security incident disclosed by LiteLLM, an open-source tool that routes requests between large language model providers, TechCrunch reported. Several other incidents were disclosed the same day.

Corma is entering a crowded field. Kevin Mandia's new start-up Armadin is also building AI for defence, and Anthropic, OpenAI and Microsoft are all pitching their own systems for security teams, alongside a long list of smaller start-ups.

By

Nakoda News Startups Desk

Founders, funding rounds, investors and the companies they back.

About the desk

Related coverage