A stealth AI cybersecurity startup named Pi, led by Elon Musk’s longtime cybersecurity expert Yoni Ramon, is launching with a $100 million valuation to build an AI security brain that prioritizes and patches vulnerabilities at scale.

Pi’s security brain learns from a client’s entire environment—incidents, policies, code, and communications—enabling rapid vulnerability assessment and prioritized fixes, including monitoring developers’ work for security issues.

Early user feedback from Navan’s CISO suggests Pi dramatically speeds up patching, enabling automatic fixes and potentially reducing the need for additional security staff by one to two full-time equivalents.

Pi is already serving high-profile clients, including a security engagement with xAI and its Grok bot on X, as well as Colossus, with a compute deal for Anthropic reportedly worth up to $1.25 billion per month through May 2029.

Pi’s leadership includes Ramon as chief product officer and ex-Microsoft security researcher Guy Arazi as CEO, with $35 million raised in a Round led by Brightmind Partners and Third Point Ventures, alongside participation from CrowdStrike’s George Kurtz and Armis cofounders.

Pi differentiates itself from rivals like Depthfirst by acting as an eidetic memory of a company’s security history, helping prevent repeating past mistakes.

The Pi platform can ingest and analyze a client’s codebase, infrastructure, and workflows to identify security risks and can process information from Slack, email, and other channels within hours, scalable to any organization.