Irregular Raises $80 Million to Secure Frontier AI Models
AI security firm Irregular raises $80 million to secure frontier AI models, marking a major milestone in the growing race to keep advanced AI safe. The funding round, led by Sequoia Capital and Redpoint Ventures, also saw participation from Wiz CEO Assaf Rappaport. With this investment, Irregular is now valued at an impressive $450 million.
Why This Funding Matters for AI Security
Co-founder Dan Lahav explained that the rapid shift toward human-on-AI and AI-on-AI interaction will challenge today’s security systems. As AI becomes more integrated into economic activity, Irregular sees an urgent need to build defenses against vulnerabilities that current tools can’t handle.
From Pattern Labs to Industry Leader
Formerly known as Pattern Labs, Irregular has become a trusted name in AI evaluations and safety testing. Its research is already shaping security reviews for major AI models, including Anthropic’s Claude 3.7 Sonnet and OpenAI’s o3 and o4-mini. The company’s SOLVE framework, used to score AI models on vulnerability detection, has gained traction across the industry.
Preparing for Emerging AI Risks
While much of its past work has focused on existing risks, Irregular is now targeting emergent behaviors that could surface unpredictably. The team is building simulated environments designed to spot harmful tendencies before they appear in real-world deployments. This proactive approach reflects a shift in AI security: from reacting to known threats toward anticipating unknown ones.
A Growing Market for AI Safety Solutions
With regulators, enterprises, and research labs all worried about AI risks, demand for frontier model security is booming. By raising $80 million, Irregular positions itself as a critical player in defining how AI safety frameworks evolve—especially as companies push AI into more sensitive and high-stakes areas.
Post a Comment