Corma Raises $60M to Build Defensive AI as Tests Show Attackers Win 88% of Simulations

Stopping a cyberattack while walking your dog - defensive AI security CEO says it's not ruff to do

Corma CEO Alon Pluda says his AI security startup aims to close the "defense gap," where models are better at offensive security. He tells the story of one customer, a security executive who was walking his dog when he received a notification on his watch from a Corma agent. "It said, 'I just caught a live attack. I need your permission to block it,'" Pluda told The Register in an interview. The security boss approved the agent's action; the agent blocked the malware and the attacker from moving across the company's network and mitigated the intrusion in under 10 minutes, Pluda said. The customer later described "walking outside with his dog, and blocking a real-live attack with his AI coworker" as "one of the most magical moments of his year," Pluda recalled. Pluda founded Corma about a year ago. And yes, all you Lord of the Rings nerds, the company gets its name from the Elven word for "ring." "We're building the one ring to rule them all, but this time for the defenders to have this power." Earlier this week, the company announced $60 million in seed funding led by Sequoia Capital, alongside Khosla Ventures and Coatue. He told us that his startup is working with Fortune 100 companies, and training models to achieve "superintelligence for defensive cybersecurity." Models from OpenAI, Anthropic, and Google are "amazingly good" at coding and language, and this includes finding and fixing bugs, and orchestrating tools across multi-step workflows, he explained. "When you combine it with agentic capabilities, they move from being incredible vulnerability researchers to end-to-end attackers," Pluda said. "So inherently, what we've seen in the last few months is the models getting exponentially better at offensive security, like we saw with the OpenAI and Hugging Face incident." But these same models aren't as skilled at carrying out defensive security tasks that don't involve scanning code for vulnerabilities and misconfigurations, he said. "The vast majority of defensive security tasks don't have anything to do with code." Corma recently tested four frontier models - Claude Opus 4.8, GPT-5.5, Grok 4.3, and DeepSeek V4 - as both attackers and defenders across the same fake company and its networks, built to closely mirror a multi-business enterprise. The attacker's task was to plant a backdoor and the defender's task was to find it and stop the attack. Closing the defensive gap Corma ran all four models against each other in every attacker and defender pairing, including each model against itself, with 15 independent engagements per pairing for 241 scored engagements. Across all of these, the models successfully implanted a persistent backdoor in 85 percent of their runs. However, these same models only detected 19 percent of attacks. "That speaks to the inherent imbalance we are trying to solve," Pluda said. "The general foundation models are getting exponentially better at offensive security, but haven't been able to improve on the same rate on defensive security. So our mission is to close this gap, and make sure the defenders win in this intelligence-versus-intelligence game - or war." Corma calls this the defensive gap, and says it has to do with the data these models are trained on and the objectives they are trained against, which lend themselves to offensive security. Defensive security, however, involves reading logs, events, configurations, audit trails, and on-disk state. This is "structured machine data that is neither prose nor source, and a small share of what these models see in training," according to Corma's research. "They appear to read it less reliably." Plus, defensive reasoning is more open-ended, while offense has a straightforward goal - like "make this work" or "break this" and a checkable finish. Agentic defenders "Defensive security," according to Pluda, "is about finding needles in the haystack." Corma's models power its AI agents, which organizations can deploy like "team members" who then operate across defensive security tasks. "It's a generalized workforce, and you can assign it to whatever security tasks you want." Fortune 100 and 500 organizations across healthcare, financial services, energy, critical infrastructure, retail, and other sectors have deployed Corma's AI workforce across their environments, according to the startup. These early deployments, we're told, have reduced threat response times by more than 94 percent, expanded security coverage by 15 times across different security functions, and uncovered multi-stage attack campaigns. "If you can get AI that is smart enough, intelligent enough, knows the domain enough, optimizes for the right things enough, and you can actually trust it, end to end, all the way to responding to real-live attacks, you can reduce all of these metrics significantly," Pluda said. "And you can cover way more ground than what is possible with just human intelligence."®

Corma raised $60M from Sequoia to build the defensive AI that cybersecurity is missing

Corma raised $60M seed from Sequoia, Khosla, Coatue. Building first defensive cybersecurity foundation model. In simulations, AI attackers won 88%, defenders caught 12%. Already deployed at Fortune 100/500 orgs. Team from DeepMind and Unit 8200. In hundreds of simulations modelled on Fortune 500 companies with dozens of security tools, Corma tested leading AI models including GPT and Claude. First, the models attacked the simulated organisations and planted persistent threats. Then the same models were asked to defend and find what they had planted. The attackers succeeded in 88% of simulations. The defenders caught 12%. The same AI that is increasingly capable of sophisticated attacks is poorly equipped to stop them. Corma raised $60 million in a seed round led by Sequoia Capital, with Khosla Ventures and Coatue, to build the defensive model the industry does not have. The company, founded in 2025 and headquartered in Tel Aviv and San Francisco, is building a foundation model from the ground up for cybersecurity defence. Rather than selling software, Corma deploys AI agents that operate across an organisation's existing security tools and carry out tasks end to end. "We don't replace anyone and we are not a product," CEO Alon Pluda told Calcalist. "We sell virtual human resources." Each organisation determines how many it needs. In early deployments at Fortune 100 and Fortune 500 companies across healthcare, financial services, energy, and retail, Corma's systems reduced threat response times by more than 94% and expanded security coverage 15x. The team brings together frontier AI researchers from Google and DeepMind with cybersecurity specialists from Israel's Unit 8200. OpenAI just paused work on its Astra model because it could not rule out that the system had reached "critical cybersecurity" capabilities, meaning it could find and exploit zero-day vulnerabilities without human help. That is the offensive side accelerating. Corma is betting that the defensive side needs its own purpose-built model rather than hoping general-purpose AI will do both jobs. The gap Corma identified is structural. Offensive cybersecurity leverages the same coding and reasoning capabilities that make frontier models powerful. Defensive cybersecurity requires processing enormous volumes of audit logs, identifying weak signals over long periods, and maintaining consistency across thousands of decisions, tasks that general-purpose models are not trained for. AI models from OpenAI, Anthropic, and Meta have all escaped test environments and breached other organisations in recent weeks. The attackers are already autonomous. Corma's argument is that the defenders need to be too.

Exclusive: Corma raises $60M from Sequoia, Khosla Ventures for AI trained to defend against cyberattacks | Fortune

The growing number of powerful, widely available AI models has ushered in a dangerous era for cybersecurity. Nefarious actors now have the ability to carry out complex attacks at unprecedented scale, creating a need for more tools to defend against them. That's where Corma comes in, a startup emerging from stealth today with $60 million in seed funding to build AI models for defensive cybersecurity, led by Sequoia Capital, along with Khosla Ventures and Coatue. The announcement did not include Corma's total valuation. Founded in 2025, Corma is based in Tel Aviv and San Francisco. It deployed its first model six weeks ago to a variety of Fortune 100 and Fortune 500 organizations across sectors, including healthcare, financial services, energy, critical infrastructure, and retail. The name "Corma" comes from The Lord of the Rings, CEO Alon Pluda tells Fortune. The books are centered around the quest to destroy an all-powerful ring, or "corma" in Elvish. Pluda says his company's product is like that ring, "but this time for the defenders." Most popular AI models today from companies like OpenAI, Anthropic, and Google are trained to carry out cybersecurity attacks, Pluda said, not to defend against them. For example, these models excel at writing and refining code, identifying bugs, and multi-step reasoning. "Those same capabilities map directly to offensive security," Pluda said. In contrast, Corma trains its AI models to specialize in qualities related to defensive cybersecurity, most of which "doesn't have anything to do with coding," Pluda said. Instead, it's more about "looking at logs, audits, [and] finding the needle in a haystack." Also key: maintaining a consistent approach across thousands of actions. "Corma has trained its model for the complexity of real-world attacks and is building the intelligence layer defense actually needs," said Shaun Maguire, Partner at Sequoia. "Agentic AI gives attackers a structural speed advantage, but Alon's hacking talent paired with Corma's frontier AI research helps companies combat these threats at scale." Corma's model has reduced threat response times by 94% in the organizations that have adopted it, the company says. As we've seen in recent months with OpenAI's models attacking Hugging Face, AI cyberattacks are happening at superhuman speed, making it difficult for human-driven processes to stop them. Corma is helping organizations fight AI with AI. "As cyberattacks become more autonomous and scalable, the stakes move beyond data and finances to critical infrastructure, healthcare systems, and essential services where failures can have real-world consequences," said Vinod Khosla, founder of Khosla Ventures. "That is why we need entirely new approaches to cyber defense and teams like Corma pursuing one of the hardest problems in cybersecurity." "Corma has trained its model for the complexity of real-world attacks and is building the intelligence layer defense actually needs," said Shaun Maguire, Partner at Sequoia. "Agentic AI gives attackers a structural speed advantage, but Alon's hacking talent paired with Corma's frontier AI research helps companies combat these threats at scale." Corma will use the $60 million in funding to further scale its AI models by expanding the data and training that power them. The company will also grow its team, with a focus on hiring top talent across defensive security, AI, and research to continue advancing their model.

Corma launches with $60M in funding for defensive cybersecurity AI

Defensive cybersecurity startup Corma Labs Ltd. today announced it has raised $60 million in seed funding to build a foundation model purpose-built for security defense. Founded in 2025, Corma runs offices in Tel Aviv and San Francisco and describes itself as a frontier artificial intelligence lab working only on the defensive side of security. Its pitch rests on a gap the company says has opened up over the past few years. Frontier models have become very good at code reasoning. Vulnerability research and exploit development are code-reasoning problems at heart, so those gains transfer directly to attackers. Defense asks for something else. Analysts sift through audit logs, events and network flows. They hold weak signals together over weeks, then make thousands of decisions in sequence without drifting. Corma ran hundreds of simulations to measure the gap. Its test environments were modeled on Fortune 500 networks, complete with the dozens of security tools a large enterprise typically runs. Leading models from OpenAI Group PBC and Anthropic PBC were first told to act as attackers and plant persistent threats inside those systems. The same models were then asked to find and remove what they had planted. AI attackers succeeded 88% of the time. The defenders caught 12%. The scenario is not hypothetical. Anthropic disclosed in July that models including Mythos 5 escaped their test sandboxes and compromised two real organizations after a configuration error left environments meant to be isolated connected to the internet. "The race to general intelligence in cybersecurity has already begun, and the attackers have a significant head start," said co-founder and Chief Executive Alon Pluda. "AI-powered attacks are operating at a speed and sophistication that neither human teams, better tooling, nor general-purpose AI can match." Corma's model powers a set of agents that customers onboard roughly the way they would a new hire. The agents work across existing security tooling rather than replacing it, and the company says they keep learning the environment they sit in. Deployments started six weeks ago. Corma says its agents are now running at Fortune 100 and Fortune 500 companies in healthcare, financial services, energy, critical infrastructure and retail. Early customers have cut threat response times by more than 94%, according to the company, and stretched coverage 15 times across security functions. Corma's team mixes AI researchers out of Google LLC and Google DeepMind with veterans of Israel's Unit 8200 and large cybersecurity vendors. Sequoia Capital Operations led the round, with Khosla Ventures and Coatue Management also taking part. Sequoia partner Shaun Maguire said Corma is "building the intelligence layer defense actually needs." Vinod Khosla, founder of Khosla Ventures, put the stakes in national security terms, arguing that autonomous attacks threaten critical infrastructure and health systems rather than just data and money. The seed round is Corma's first disclosed outside funding.

Corma, the First Frontier Defensive Cybersecurity AI Lab, Raises $60M as AI Supercharges Attackers

Backed by Sequoia Capital and Khosla Ventures, Corma is building the first foundation model for defensive cybersecurity, closing the growing asymmetry between AI-powered offense and human-led defense. SAN FRANCISCO, August 10, 2026 (Newswire.com) - Corma, the first frontier AI lab for defensive cybersecurity, today announced $60 million in seed funding led by Sequoia Capital, alongside Khosla Ventures and Coatue. Already working with Fortune 100 companies, Corma is developing the first foundation model purpose-built for defensive cybersecurity, tackling a rapidly growing challenge the world is only beginning to confront: AI's offensive cybersecurity capabilities are advancing at an extraordinary pace, while its defensive cybersecurity capabilities continue to fall short of protecting organizations in real-world environments. Foundation models such as OpenAI's GPT, Anthropic's Claude, and Google's Gemini have advanced at an extraordinary pace over the past few years, particularly in coding and software reasoning. These systems can now write and refine software, identify and remediate bugs, reason through complex environments, and orchestrate tools across multi-step workflows. Those same capabilities map directly to offensive security. Vulnerability research and exploit development are, at their core, code-reasoning problems executed against bounded targets. When combined with agentic execution, these models move beyond assisting attackers to autonomously carrying out end-to-end attack chains, as demonstrated in Anthropic's Mythos disclosure. But defensive cybersecurity - everything outside of code security - demands a fundamentally different set of capabilities. Rather than code reasoning and generation, it requires digging through massive volumes of security data (such as audit logs, events, and network flows), correlating weak signals over long time horizons, and maintaining extreme consistency across thousands of decisions in sequence. Corma's research shows just that. Using realistic enterprise environments modeled after Fortune 500 organizations, complete with the dozens of security tools typically deployed in large enterprises, Corma ran hundreds of simulations with leading AI models such as OpenAI's GPT and Anthropic's Claude. First, the models were tasked with acting as attackers, planting persistent threats inside enterprise systems. The same models were then asked to defend those environments, locating and remediating the threats they had created. The results were consistent: in nearly every case, the same AI model that executed an end-to-end attack couldn't defend against the very attack it carried out - AI attackers succeeded 88% of the time, while AI defenders detected just 12%. This highlights a growing imbalance with major implications for cybersecurity. "The race to general intelligence in cybersecurity has already begun, and the attackers have a significant head start," said Alon Pluda, Co-founder and CEO of Corma. "AI-powered attacks are operating at a speed and sophistication that neither human teams, better tooling, nor general-purpose AI can match. It requires a complete AI-powered defensive workforce, built from the ground up for cybersecurity, that gives defenders the same speed, sophistication, and generalization that AI has already given attackers. Corma's mission is to make sure the defenders win this race - and every challenge that comes next." Corma is building the first foundation model purpose-built for defensive cybersecurity. It powers Corma's AI agents, allowing them to outperform agents built on other models while generalizing across the full spectrum of defensive security tasks. Organizations onboard Corma much like they would a new team member. Once deployed, Corma's agents can operate across virtually every area of defensive cybersecurity, continuously learn the environment around them, and scale to meet demands that no team could cover alone. Since launching just six weeks ago, Corma's AI workforce has been deployed at Fortune 100 and Fortune 500 organizations across healthcare, financial services, energy, critical infrastructure, retail, and other sectors. Early deployments have reduced threat response times by more than 94%, expanded security coverage by 15 times across different security functions, and uncovered multi-stage attack campaigns that would have otherwise gone undetected. "Corma has trained its model for the complexity of real-world attacks and is building the intelligence layer defense actually needs," said Shaun Maguire, Partner at Sequoia. "Agentic AI gives attackers a structural speed advantage, but Alon's hacking talent paired with Corma's frontier AI research helps companies combat these threats at scale." "AI is reshaping cybersecurity in ways that increasingly extend beyond the enterprise to national security and geopolitical stability," said Vinod Khosla, founder of Khosla Ventures. "As cyberattacks become more autonomous and scalable, the stakes move beyond data and finances to critical infrastructure, healthcare systems, and essential services where failures can have real-world consequences. That is why we need entirely new approaches to cyber defense and teams like Corma pursuing one of the hardest problems in cybersecurity." Corma brings together a first-of-its-kind combination of expertise across pre-training, post-training, offensive and defensive cybersecurity - uniting frontier AI researchers from Google and DeepMind with cybersecurity experts from the most elite groups within Israel's 8200 Unit and the world's leading cybersecurity companies. It is the type of interdisciplinary team required to tackle Corma's mission, long considered "the holy grail of cybersecurity." About Corma Corma is the first frontier AI lab for defensive cybersecurity. The company is building the first foundation model purpose-built for defensive cybersecurity, powering a new kind of AI native defensive workforce that gives security teams the same speed, sophistication, and generalization that AI has already given attackers. In a world where AI-powered attacks are advancing at unprecedented speed, Corma's mission is to make sure the defenders win the race - and every challenge that comes next. Founded in 2025 and headquartered in Tel Aviv and San Francisco, Corma is backed by Sequoia Capital, Khosla Ventures, and Coatue."