A $60 million seed round is a statement. When it comes from Sequoia Capital and Khosla Ventures, it is a thesis. The thesis, backing a company called Corma that emerged from stealth in August 2026, is that the next major cybersecurity workforce will not be human [Fortune, Aug 2026].
The company is training foundation models specifically for defensive cybersecurity, aiming to deploy them as a "virtual security workforce" of AI agents [SiliconANGLE, Aug 2026]. These agents are designed to work inside a customer's existing security tools, detecting, investigating, and responding to threats. The bet is not on a new dashboard but on a new kind of employee, one that generalizes across tasks and never logs off.
The Defensive Gap Thesis
Corma's founding premise is what it calls the "defense gap." The company argues that while AI has rapidly scaled offensive cyber capabilities, general-purpose models are not optimized for the nuanced work of defense [govinfosecurity.com, Aug 2026]. Its solution is to build a specialized foundation model from the ground up, trained on defensive security tasks.
The product vision is operational, not theoretical. The AI agents are meant to be onboarded like a new team member, integrating with the security information and event management (SIEM) systems, endpoint detection tools, and ticketing platforms already in use [SiliconANGLE, Aug 2026]. The promise is to augment human teams by taking on the volume of alerts and initial investigation, allowing analysts to focus on higher-order judgment.
A Team Forged in Intelligence and AI
The credibility for such an ambitious technical bet rests on the team. Corma's founders and early hires pull from two elite pools: top-tier AI research and Israeli military intelligence.
- AI Research. The team includes researchers who are alumni of Google and Google DeepMind, bringing experience in building large-scale machine learning systems [Calcalistech, Unknown].
- Cybersecurity Operations. The other wing of the team is built from veterans of Israel's Unit 8200, the renowned signals intelligence unit, and specialists from major cybersecurity vendors [Calcalistech, Unknown].
This blend is deliberate. It pairs the builders of advanced AI with operators who understand the reality of cyber defense. CEO and co-founder Alon Pluda is a longtime member of Israeli Military Intelligence, according to public profiles [govinfosecurity.com, Aug 2026].
| Role | Name | Noted Background |
|---|---|---|
| CEO & Co-Founder | Alon Pluda | Israeli Military Intelligence [Crunchbase, retrieved 2026] |
| Co-Founder | Dana Raveh Tiram | VP Marketing [Startup Nation Finder, retrieved 2026] |
| Co-Founder | Yuval Zacharia | Announced company launch [LinkedIn, Yuval Zacharia, retrieved 2026] |
| Co-Founder | Nitzan Simchi | AI & Algorithms Researcher [LinkedIn, Nitzan Simchi, retrieved 2026] |
The Capital Stack and the Clock
The $60 million seed, led by Sequoia Capital with participation from Khosla Ventures and Coatue, is a war chest [Fortune, Aug 2026]. It is also a timer. Building a foundational model for cybersecurity is a capital-intensive, multi-year endeavor requiring top-tier research talent and massive compute resources. This round finances that runway.
For the investors, the logic is clear. Sequoia has publicly stated that Corma is training a foundation model for enterprises to defend themselves against AI’s scaling offensive capabilities [Sequoia Capital, Unknown]. In a market where security talent is perpetually scarce and attack surfaces are exploding, an effective AI multiplier has obvious enterprise value.
Seed Round (Aug 2026) | 60 | M USD
Where the Wheels Could Come Off
The ambition is matched by significant technical and market risks. Corma is entering a field crowded with established security vendors adding AI features and a growing cohort of AI-native security startups. Its differentiation hinges entirely on the performance of its proprietary foundation model, which remains unproven at scale.
The core challenges are threefold.
- Model Generalization. The promise is a workforce that generalizes across defensive tasks. Achieving true generalization in the messy, context-dependent world of security operations is an unsolved problem. A model that is merely good at several tasks is not the same as a generalized analyst.
- Integration Depth. Successful deployment "where the human security team already works" requires deep, reliable integrations with a vast and heterogeneous enterprise toolset [corma.ai, retrieved 2026]. This is a heavy lift in engineering and partnership development.
- The Proof Gap. The company emerged with funding but without public metrics on customers, response times, or false-positive rates. Enterprise security buyers are notoriously conservative; they will require concrete evidence of efficacy before trusting an AI with their crown jewels.
Corma's answer to these risks is its team composition and its capital. The blend of AI and operational expertise is designed to build a model that understands real-world security, not just academic benchmarks. The $60 million provides the time to do it right, away from the pressure of immediate revenue.
The Next Twelve Months
The coming year will be about moving from thesis to proof. The key milestones to watch will be the announcement of design partners or early customers, the publication of any technical research demonstrating the model's capabilities, and the expansion of the team, particularly in enterprise sales and security engineering.
The seed round from Sequoia, Khosla, and Coatue sets a high bar. It values the team's pedigree and the strategic importance of the problem. The question for the market now is whether Corma can translate that $60 million vote of confidence into a product that security chiefs are willing to hire.
Sources
- [Fortune, Aug 2026] Exclusive: Corma raises $60M from Sequoia, Khosla Ventures for AI trained to defend against cyberattacks | https://fortune.com/2026/08/10/exclusive-corma-raises-60-million-from-sequoia-for-ai-trained-to-defend-against-cyberattacks/
- [SiliconANGLE, Aug 2026] Corma launches with $60M funding for defensive cybersecurity AI | https://siliconangle.com/2026/08/10/corma-launches-60m-funding-defensive-cybersecurity-ai/
- [The Register, Aug 2026] Stopping a cyberattack while walking your dog? Defensive AI security CEO says it's not ruff to do | https://www.theregister.com/security/2026/08/16/stopping-a_cyberattack-while-walking-your-dog-defensive-ai-security-ceo-says-its-not-ruff-to-do/
- [Calcalistech, Unknown] Corma, an Israeli-American cybersecurity startup, raises $60 million in seed funding | https://www.calcalistech.com/ctechnews/article/h1cwzhvigg
- [govinfosecurity.com, Aug 2026] Corma says general-purpose models aren't optimized for defensive security | https://govinfosecurity.com
- [Sequoia Capital, Unknown] Corma is training a foundation model for defensive cybersecurity | https://sequoiacap.com
- [Crunchbase, retrieved 2026] Alon Pluda profile | https://crunchbase.com
- [Startup Nation Finder, retrieved 2026] Dana Raveh Tiram profile | https://finder.startupnationcentral.org
- [LinkedIn, Yuval Zacharia, retrieved 2026] Yuval Zacharia profile | https://il.linkedin.com
- [LinkedIn, Nitzan Simchi, retrieved 2026] Nitzan Simchi profile | https://il.linkedin.com
- [corma.ai, retrieved 2026] Corma company claims | https://corma.ai/