Thursday, November 19, 2026 Updated Briefings Saved 中文
AI Daily Radar

AI and the frontier tech around it, explained every morning

Announced

Corma raises $125M Series A for cyber-defense AI agents

Photo: cliff1126 / Pixabay

Corma, a year-old startup with offices in Tel Aviv and San Francisco, raised a $125 million Series A led by NEA, according to a Wall Street Journal report by Angus Loten. The round follows a $60 million seed the company raised in August, meaning Corma has disclosed roughly $185 million in funding within about three months of launching publicly. Corma builds AI agents that it says are custom-trained for defensive cybersecurity — software that watches for and responds to attacks rather than generating them. The report excerpt does not give a valuation, named customers, or say whether the agents are generally available, so the product’s real-world deployment remains unverified.

Verified and unconfirmed

  • Verified: The Wall Street Journal’s Angus Loten reports that Corma raised a $125 million Series A led by NEA, after a $60 million seed round in August, and describes the company as a year-old, Tel Aviv- and San Francisco-based builder of AI agents custom-trained for defensive cybersecurity.
  • Not yet confirmed: The company’s valuation, its customers and revenue, how many agents it has deployed, and whether the product is generally available — none of these appear in the source. The description of what the agents do rests on the company’s own characterization, and no independent test or benchmark of the agents is cited.

Background

Corma was founded in 2025 by CEO Alon Pluda, a Unit 8200 alumnus, and operates from Tel Aviv and San Francisco with roughly 25 employees. It describes its core technology as a foundation model purpose-built for defensive cybersecurity, aimed at the mismatch between AI-powered attackers and human-led defenders.

Why it matters

With about $185 million disclosed across two rounds in roughly three months, Corma is among the most heavily funded young entrants in security AI, giving it room to hire research and engineering staff and to sell into security-operations budgets that incumbent vendors also target. For security teams evaluating such tools, the practical hurdle is unchanged: “custom-trained” agents need to be tested against the buyer’s own telemetry and alert data before replacing or augmenting existing detection and response workflows.

What to watch

Watch for Corma’s first named customer deployments or any independent benchmark results for its agents, since the funding report provides neither; a further round or a valuation disclosure would also show how investors are pricing security-agent startups.

Further reading

Comments

Comments

Comments use a GitHub account and are moderated. Comment rules

More in AI

See all