The Irregular AI security startup has secured $80 million in fresh funding to tackle emerging threats in frontier AI models. The round, which values the company at $450 million, was jointly led by Sequoia Capital and Redpoint Ventures, with participation from Wiz CEO Assaf Rappaport. The new capital will be used to build complex simulated environments where AI models can be rigorously tested before their public release.
Irregular, originally known as Pattern Labs and it has become a key player in AI safety assessments. The firm’s proprietary scoring framework ‘SOLVE’, is now an industry-wide tool for evaluating model vulnerabilities. Its work has been instrumental in the security evaluations for prominent models, including Anthropic’s Claude 3.7 Sonnet and OpenAI’s o3 and o4-mini models.
Also See: Jaguar Land Rover Extends Production Shutdown to a Third Week Following Crippling Cyberattack
Co-founder Dan Lahav noted that the firm is shifting its focus toward identifying “emergent risks”—unpredictable security flaws that only appear after models are deployed. This proactive approach is central to the mission of Irregular AI security: to get ahead of threats by using sophisticated simulations where AI agents take on roles as both attackers and defenders, flagging weaknesses that might otherwise be missed.
This focus comes as security concerns intensify across the industry. As large language models become more capable of independently finding and exploiting software vulnerabilities, the stakes are rising. OpenAI recently overhauled its own security measures to prevent corporate espionage, highlighting the growing need for specialized solutions like those developed by Irregular AI security.
With this new funding, Irregular is positioning itself at the forefront of a critical niche that could shape the future of AI safety management. The strong investor confidence signals a clear market demand for their approach, underscoring that as AI’s impact grows, so does the urgency for robust security protocols.
You may also like: OpenAI Corporate Restructuring: Microsoft Deal Aims for Public Benefit Status
