{
  "id": 7652777,
  "title": "AI agent certification startup AIUC raises $40M to begin auditing frontier models",
  "url": "https://urgent.news/2026/09/15/ai-agent-certification-startup-aiuc-raises-40m-to-begin-auditing",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-15T22:44:14.000Z",
  "source": {
    "name": "SiliconANGLE",
    "slug": "siliconangle",
    "url": "https://siliconangle.com/2026/09/15/ai-agent-certification-startup-aiuc-raises-40m-to-begin-auditing-frontier-models/"
  },
  "original_language": "en",
  "account": "AI Underwriting Company, a startup specializing in certifying artificial intelligence agents, has recently secured $40 million in funding to commence auditing frontier AI models. Currently, their efforts have focused solely on agents built on these models, but the company believes that risk uncertainty is now a more significant hindrance to AI adoption than any limitations in the technology's capabilities. Rune Kvist, the co-founder and CEO of AIUC, explained that many enterprises have approved agents in pilot programs but are stalled at the security review stage. He added that obtaining proof of security and reliability has become the primary bottleneck.\n\nThe company's standard for certification, called AIUC-1, involves subjecting an agent to approximately 5,000 risk and attack combinations that pertain to the specific business deploying it. These tests include scenarios such as jailbreaks, hallucinations, and data leaks, among others. Each agent undergoes independent auditing and recertification every quarter as new attack techniques emerge. Many of these tests are performed by AI agents, with human verification of the final audit.\n\nSeveral companies, including Eleven Labs Inc., Anysphere Inc., Harvey AI Corp., KPMG LLP, Lovable Labs Inc., UiPath Inc., and Fin (formerly Intercom Inc.), have already obtained certification. The certification process is guided by a consortium of over 250 security and risk leaders from Fortune 1000 companies who advocate for its adoption within their organizations. Kvist and his co-founder, Rajiv Dattani, joined AIUC from different perspectives. Kvist was Anthropic PBC's first product hire, while Dattani worked as a partner in McKinsey & Co.'s insurance practice and later served as the chief operating officer of AI model evaluation nonprofit METR.\n\nThe founders liken the situation to the early days of electrical safety, when insurers funded Underwriters Laboratories to test and certify products, and the UL mark became ubiquitous on light bulbs in America. Dattani noted that, just as insurance funded safety standards for electrical products, AI needs a similar combination of standards, testing, and insurance to ensure safety and reliability. The company started by targeting applications since security reviews were blocking rollouts there, but now, the same pressure is growing around frontier models. The funds raised will expand the company's audit and insurance work to cover these advanced AI systems, which the company describes as an ecosystem for frontier oversight. Ribbit Capital led the Series A funding round with participation from First Harmonic, and Nat Friedman at NFDG led a $15 million seed round when the company launched in July 2025. The total investment in AIUC now stands at $55 million. Micky Malka, founder and managing partner of Ribbit Capital, remarked that his firm has been backing financial services companies for a decade, where trust is paramount, and AI is now entering that same realm of trust-building, moving faster than existing evaluation systems.",
  "summary": "Artificial intelligence agent certification startup Artificial Intelligence Underwriting Company today announced it has raised $40 million in new funding to begin auditing frontier AI models. Until now, its work has covered only the agents that companies build on those models. AIUC argues that uncertainty over risk now slows AI adoption more than any shortfall in what […] The post AI agent…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}