AI Security Startup Irregular Raises $80M to Fortify AI Safety

Irregular, an emerging leader in AI security, has secured $80 million in a funding round led by Sequoia Capital and Redpoint Ventures, with participation from Wiz CEO Assaf Rappaport. The investment values the company at $450 million, underscoring growing investor confidence in AI safety solutions.

Image source note: The image is AI-generated, and the image licensing service is Midjourney

Pioneering AI Risk Assessment

Founded as Pattern Labs, Irregular has established itself as a key player in AI model evaluation. Its frameworks are routinely used to assess safety in major models like Claude3.7Sonnet and OpenAI's o3/o4-mini. The company's SOLVE scoring system for model vulnerability detection has become an industry standard.

"We're preparing for an economy where most transactions will involve human-AI or AI-AI interactions," explained co-founder Dan Lahav. "These will fundamentally challenge existing security paradigms."

Simulating Tomorrow's Threats Today

The new funding will accelerate development of Irregular's proprietary simulation environments that stress-test AI models before deployment. These complex networks pit AIs against each other as both attackers and defenders, revealing vulnerabilities under controlled conditions.

"Our simulations show where defenses hold and where they collapse," said co-founder Omer Nevo. "This proactive approach lets us identify novel risks before models reach production."

The Expanding Frontier of AI Security

The investment comes amid heightened focus on AI safety across the industry:

• OpenAI recently overhauled internal security to prevent corporate espionage risks
• Advanced models now demonstrate unprecedented capability to discover software vulnerabilities
• Governments are implementing stricter regulations around high-risk AI applications

"If labs keep building more powerful models, we'll keep developing ways to secure them," Lahav noted. "It's an endless race, but one that's critical for responsible AI development."

Key Points:

• 💰 $80M funding round at $450M valuation
• 🔍 Specializes in pre-deployment risk identification
• ⚔️ Uses adversarial simulations to test model defenses
• 🛡️ SOLVE framework widely adopted for vulnerability scoring
• 🌐 Addressing emerging threats from human-AI/AI-AI interactions