- $1 billion valuation achieved in just 90 days
- $125 million Series A funding round
- Autonomous AI workforce reduces security task costs by over 90%
Experts would likely conclude that Corma's rapid rise reflects a critical market need for AI-driven defensive cybersecurity to counter the growing asymmetry between offensive and defensive AI capabilities.
The 90-Day Unicorn: Inside Corma's $1 Billion Bet on Defensive AI
SAN FRANCISCO – November 16, 2026 – In the traditional rhythms of venture capital, the journey from a stealth seed round to a billion-dollar valuation is a multi-year marathon requiring sustained revenue growth, market testing, and iterative product development. For Corma, a frontier artificial intelligence lab focused exclusively on defensive cybersecurity, that journey took exactly 90 days.
Today, Corma announced a $125 million Series A funding round, catapulting the company to a $1 billion valuation just three months after its initial launch. Led by NEA, with participation from existing heavyweight investors Sequoia Capital, Khosla Ventures, and Coatue, the round brings Corma's total capital raised to $185 million.
But behind the eye-watering valuation and the compressed timeline lies a much larger structural story about the 21st-century marketplace. We are witnessing a fundamental, terrifying asymmetry in the development of artificial intelligence—one that has left the world's largest enterprises dangerously exposed. The rapid ascent of Corma is not merely a symptom of venture capital exuberance; it is a desperate, highly capitalized scramble to build the infrastructure necessary to survive an era where cyber warfare is entirely autonomous.
The Offense-Defense Asymmetry in the Age of AI
To understand the strategic rationale behind Corma's explosive growth, one must first understand a core architectural flaw in how today's frontier AI models are built. For the past several years, the central race among the world's leading AI labs has been coding proficiency. Generative models have been trained on vast repositories of the world's software, learning to write, debug, and optimize code at an unprecedented scale.
However, the AI that is learning to write the world's software is simultaneously learning how to break it. In the realm of computer science, a vulnerability is simply another type of bug. As general-purpose large language models become exponentially better at reasoning through complex codebases, they inherently become exponentially better at discovering zero-day vulnerabilities. When paired with autonomous agents, these models can exploit what they find from end to end, operating at a speed that human security teams cannot match.
This dynamic creates a massive structural advantage for attackers. General-purpose AI models are essentially acting as automated vulnerability discovery engines. Internal industry simulations have starkly illustrated this defensive gap: when general-purpose AI models are tasked with executing cyberattacks, they achieve staggering success rates, sometimes breaching simulated environments nearly 90 percent of the time. Conversely, when those exact same models are repurposed for defensive tasks, their success rate plummets to the low double digits.
Defensive cybersecurity demands a fundamentally different set of capabilities. It is not an exercise in code reasoning; it is an exercise in signal processing. Defense requires sifting through massive, unstructured volumes of security data—audit logs, endpoint events, and network flows. It requires correlating weak, seemingly unrelated signals across disparate systems over long time horizons, all while filtering out enormous amounts of daily operational noise. General-purpose models, optimized for predicting the next token in a line of Python, were never designed for this kind of forensic endurance.
Black October and the Market Reality
Theoretical warnings about this asymmetry became a harsh market reality last month. In what the cybersecurity industry has grimly dubbed "Black October," a relentless wave of critical vulnerabilities and automated breaches swept across some of the world's most widely used enterprise software. The sheer volume and speed of the attacks overwhelmed traditional security operations centers, prompting frantic alerts from government cybersecurity agencies stretching from Washington to Tokyo.
"Cybersecurity is now a battle of intelligence against intelligence, and Black October made that impossible to ignore," said Alon Pluda, Co-founder and CEO of Corma. Pluda, a former operative in Israel's elite Unit 8200 military intelligence group, recognized early that human speed was no longer a viable defense against machine-speed attacks.
The fallout from Black October served as a brutal catalyst for enterprise adoption. Since emerging from stealth in August, Corma has seen a surge in demand across the Fortune 100, Fortune 500, and Fortune Global 500. Organizations in critical sectors—ranging from healthcare and financial services to energy and telecommunications—are racing to integrate Corma's foundation models into their defense stacks.
"Since we came out of stealth, demand from the world's largest organizations has been overwhelming, and one of our hardest problems has been securing capacity fast enough to keep up with the demand," Pluda noted.
The Three-Month Unicorn and Capital Frenzy
Meeting that overwhelming enterprise demand requires immense computational power. Over the past three months, Corma has been forced to scale its server and compute infrastructure by a staggering 24 times just to keep pace with customer deployments. This aggressive infrastructure expansion is the primary driver behind the $125 million Series A. Building and running specialized foundation models is a capital-intensive endeavor, and Corma is competing for the same scarce computing resources as the broader generative AI market.
The investor syndicate backing Corma represents a consensus among the tech industry's most influential kingmakers that legacy cybersecurity paradigms are obsolete. Sequoia Capital, which led Corma's $60 million seed round in August, doubled down in the Series A alongside Khosla Ventures and Coatue.
NEA's decision to lead the Series A underscores the perceived market opportunity. "Corma is building the intelligence layer that defensive security needs," said Aaron Jacobson, Partner at NEA. "As AI gives attackers an advantage, Corma's combination of frontier AI research and deep cybersecurity expertise has the potential to help organizations defend against these threats at scale. Corma is the only company taking this on in a way that is scalable across organizations to prevent the next big cyberattack."
To execute this vision, Corma has assembled a formidable talent pool. Headquartered in both Tel Aviv and San Francisco, the company has aggressively recruited researchers from Google DeepMind, pairing them with offensive security experts and veterans of the Israeli intelligence apparatus. This fusion of frontier AI research and battle-tested cyber warfare experience is the bedrock of the company's technological differentiation.
Replacing the Tier-1 Analyst: The Autonomous Workforce
The ultimate disruption Corma brings to the global business landscape is not just technological, but operational. The company does not merely sell a software tool; it deploys an autonomous AI workforce.
For decades, enterprise security has relied on a tiered human analyst model. Tier-1 analysts sit in security operations centers, triaging thousands of daily alerts generated by passive monitoring tools like firewalls and SIEMs (Security Information and Event Management systems). It is grueling, repetitive work characterized by high burnout rates and alert fatigue.
Corma's foundation models are designed to ingest this workflow entirely. Operating within an organization's existing security infrastructure, the AI workforce autonomously investigates alerts, correlates data across cloud and endpoint environments, and executes remediation steps. Early enterprise deployments suggest a paradigm shift in operational efficiency, with some early adopters reporting that the AI workforce can complete the vast majority of routine security tasks autonomously, drastically reducing end-to-end response times and driving down the cost per completed task by over 90 percent.
This shift from passive alerting to active, autonomous defense represents a structural change in how corporations manage risk. Leaders are no longer simply buying software to help their human teams work faster; they are delegating the frontline defense of their digital infrastructure directly to artificial intelligence. As the events of Black October demonstrated, the attackers have already automated their operations, leaving enterprises with little choice but to fight fire with fire, fundamentally altering the future of competition and security in the digital age.
Topics & Related
Artificial Intelligence
Agentic AI
Cybersecurity
📝 This article is still being updated
Are you a relevant expert who could contribute your opinion or insights to this article? We'd love to hear from you. We will give you full credit for your contribution.
Contribute Your Expertise →