The field of AI security has welcomed a major player. AI security company Irregular announced on Wednesday that it has completed a new round of funding of 80 million USD, led by Sequoia Capital and Redpoint Ventures, with participation from Assaf Rappaport, CEO of the cybersecurity unicorn Wiz. According to insiders, the company's valuation in this round of financing is 450 million USD.
This company, formerly known as Pattern Labs, has already taken an important position in the AI evaluation field. The research results of Irregular have been cited in the security evaluations of Claude 3.7 Sonnet and OpenAI's o3 and o4-mini models. More notably, the company's vulnerability detection capability scoring framework SOLVE has been widely applied in the industry.
Co-founder Dan Lahav explained the company's vision to TechCrunch: "We believe there will be a large amount of economic activity coming from interactions between humans and AI, as well as between AI and AI, which will break existing security systems at multiple points." This forward-looking judgment is one of the key reasons why Irregular has attracted investor favor.
Although Irregular has made significant contributions to assessing risks of existing models, the company's current financing aims for more ambitious goals: to identify emerging risks and behaviors before they appear in the real world. To achieve this, the company has built a complex simulation environment system that can conduct intensive testing before model release.
Co-founder Omer Nevo explained their testing method: "We have a complex network simulation environment, allowing AI to play both the roles of attacker and defender. When a new model is released, we can see where the defenses are effective and where they are lacking."
This innovative testing method reflects the complex challenges faced in the AI security field. As AI model capabilities continue to grow, they are becoming increasingly skilled at identifying software vulnerabilities, a capability that is significant for both attackers and defenders.
Attention to security issues in the AI industry has risen sharply. OpenAI completely reformed its internal security measures this summer, focusing on preventing potential corporate espionage. At the same time, the potential risks brought by cutting-edge AI models continue to emerge, making security assessments more important than ever.
Irregular's business model is based on a key insight: as the capabilities of large language models grow, security challenges will increase exponentially. Lahav said, "If leading laboratories aim to create increasingly complex and powerful models, our goal is to protect the security of these models. But this is a moving target, so there is still a lot of work to be done in the future."
This dynamic security challenge creates huge market opportunities for Irregular. As AI systems are widely applied across industries, the demand for professional security assessments and protection will continue to grow. The company not only needs to address current known risks but also has the ability to predict and prevent future risks.
The lead investment from Sequoia Capital and Redpoint Ventures shows the importance that top investment institutions place on the AI security field. With the rapid development of AI technology, security issues have risen from technical details to strategic challenges, requiring specialized solutions and professional teams.
The participation of Wiz's CEO in the investment also has symbolic significance, indicating that leaders in the traditional cybersecurity field have recognized the importance and market potential of AI security. AI security shares overlaps with traditional cybersecurity but also has differences, requiring new methods and tools to address.
When AI models become increasingly powerful, and human-machine interaction and AI-AI interaction become the main forms of economic activity, security guarantees will become the cornerstone of the entire AI ecosystem. The rise of Irregular is not just a company's success, but also represents the vigorous development of this emerging sector of AI security. In this era where technology and security evolve in parallel, companies like Irregular will play an increasingly important role in ensuring the healthy development of AI technology.
