Technical

Srinivasa Reddy Kandi: Irregular Raises $80 Million to Strengthen Frontier AI Security

September, 18, 2025-04:17

Share: Facebook | Twitter | Whatsapp | Linkedin | Visits: 37817 | 2821


Srinivasa Reddy Kandi: Irregular Raises $80 Million to Strengthen Frontier AI Security


Irregular Raises $80 Million to Strengthen Frontier AI Security:

AI security startup Irregular has raised $80 million in a funding round led by Sequoia Capital and Redpoint Ventures, with additional participation from Wiz CEO Assaf Rappaport. The round values the company at $450 million, according to a source familiar with the deal.

Co-founder Dan Lahav explained the company’s vision to TechCrunch: “Soon, much of the world’s economic activity will come from human-on-AI and AI-on-AI interactions — and that’s going to challenge the security stack at multiple levels.”

Formerly known as Pattern Labs, Irregular has become a key player in AI security evaluations. Its research has been cited in safety assessments for Claude 3.7 Sonnet as well as OpenAI’s o3 and o4-mini models. The company’s vulnerability-scoring framework, SOLVE, is now widely adopted across the industry.

While Irregular has built a strong reputation for analyzing existing model risks, the new funding will accelerate its push into anticipating emergent threats — risks and behaviors that may not appear until a model is deployed in real-world environments. To support this, Irregular has developed complex simulated environments where AI agents take on both attacker and defender roles, stress-testing models before release.

“When a new model comes out, we can run it through our simulations and see exactly where defenses succeed and where they fail,” said co-founder Omer Nevo.

Security is rapidly becoming a central concern for the AI industry. OpenAI recently revamped its own internal security measures amid concerns about corporate espionage, while researchers continue to warn of models’ increasing ability to identify software vulnerabilities — a capability that could benefit both defenders and malicious actors alike.

Author: Kandi Srinivasa Reddy, Srinivasa Reddy Kandi, #KandiSrinivasaReddy, #SrinivasaReddyKandi



Leave a Comment