Rein Security raises $25M Series A to guard AI agents at runtime

Rein Security, a New York/Tel Aviv-based startup, has raised a $25 million Series A co-led by Glilot Capital and Sienna Venture Capital, bringing total funding to $35 million. The news was first reported by SecurityWeek, with corroborating coverage from PRNewswire and Calcalist.

What differentiates Rein’s pitch from a growing pile of “secure the AI agent” startups is its runtime-first architecture. The company uses patented sidecar technology that sits alongside agents as they execute, giving security teams visibility into every line of code an agent runs and every resource it touches — rather than relying solely on pre-deployment scanning or static policy review. Critically for data security teams, Rein’s platform messaging explicitly frames this as data-access governance: it enforces guardrails “dynamically and in real time with business-aware controls based on user, role, and data sensitivity.” That puts Rein in the same emerging lane as Archestra and Hush Security, both of which are building controls specifically for what autonomous agents can read, write, or exfiltrate.

Calcalist reports the company’s revenue has grown eightfold since January, with the platform now securing thousands of agents executing millions of actions. Named enterprise customers reportedly include Lemonade and Dun & Bradstreet — concrete signal of production deployment rather than pilot-stage interest.

For DLP and DSPM buyers, this is another data point that “control what the agent can touch” is solidifying into its own category, wedged between classic DLP policy enforcement and identity/access governance. Incumbent DSPM and DLP vendors expanding into agent guardrails should treat this space as both a channel-partner opportunity and a competitive overlap risk — the sensitivity-aware access controls Rein is building look increasingly like a natural extension of existing data classification and policy engines, not a wholly separate discipline.