
Business
Vanta Agrees to Buy Pillar Security, Reportedly for About $150 Million, to Police AI Agents in Real Time
The compliance platform, which counts more than 20,000 customers, wants to move from documenting AI rules to enforcing them, with runtime blocking promised by summer 2027.
San Francisco · Elena Vos, Business · Nov 10 2026
SAN FRANCISCO - Vanta, the security and compliance software company used by more than 20,000 businesses, has signed a definitive agreement to acquire Pillar Security, a young company that tests AI agents and blocks them when they step out of line. Vanta announced the deal on Monday in a press release and did not disclose terms. Israeli business daily Calcalist reported that the deal is valued at about $150 million in cash and shares.
The pitch is that written policy cannot keep up with software that acts on its own. “Customers, regulators and auditors want proof that a company’s AI does what it says on the tin. A policy document doesn’t get you there,” said Christina Cacioppo, Vanta’s chief executive and co-founder. “Pillar Security lets companies test their agents and stop them when they go out of bounds, and Vanta turns all of that into evidence.”
Screenshot: @TrustVanta on X
Pillar was founded in 2023 by Ziv Karliner and Dor Sarig and sells a platform to discover, test, secure and govern AI applications and agents. According to Calcalist, it has raised $9 million, employs more than 20 people and will become Vanta’s research and development center in Israel, with both founders staying on to lead its AI security work. In a blog post, the founders said their research team had found vulnerabilities rated CVSS 10, including a sandbox escape in the n8n automation platform and a flaw in Google’s Gemini CLI, and that their SAIL framework for securing AI systems has passed 50,000 downloads.
Vanta has laid out a public schedule for folding Pillar in. An AI inventory that maps a company’s AI applications, models, agents and Model Context Protocol servers to their owners and risks arrives in January 2027, alongside runtime monitoring that flags suspicious agent activity. Red-team testing tailored to a customer’s business follows in spring 2027, and runtime enforcement, which blocks unauthorized agent actions as they happen, is due in summer 2027. Vanta said its AI Governance product will become a paid offering in January.
Screenshot: Vanta
In practice, Vanta says, that means an agent trying to open restricted data or post sensitive financial information in an unapproved channel would be stopped, and the Vanta Agent would alert the owners and log which protections fired. “An agent can cross a line in a single action, long after its last review,” Sarig said in Vanta’s announcement. “That’s why security has to live where agents act.”
The deal builds on the AI governance product Vanta introduced in July, which it says drew the largest early-access waitlist in its history. Vanta said in April that its annual recurring revenue had passed $300 million, up 63% from a year earlier, and its customer count has grown from 16,000 in the first quarter of 2026, citybiz reported. Pillar would be its third acquisition, after Trustpage in 2023 and Riskey in 2025. The transaction is still subject to closing, and Vanta plans to show more of its AI governance roadmap at its VantaCon event this week.