Eve Security Extends Seed To $7.5M To Police AI Agents At Runtime

Austin-based Eve Security has added $4.5M to a $7.5M seed round led by Run Ventures, pushing runtime observability, session tainting and deterministic policy enforcement into enterprise AI agent deployments.

Eve Security Extends Seed To $7.5M To Police AI Agents At Runtime

Eve Security has extended its seed round to $7.5 million to build a runtime security layer for enterprises deploying autonomous AI agents, saying the market has moved faster than the defensive stack meant to police it. The Austin-based startup added $4.5 million in new capital led by Run Ventures, with Dreamit Ventures, Blu Ventures and returning backer LiveOak Ventures joining the round announced on September 15, 2026.

Why runtime security keeps eating budget

Enterprise agent frameworks now sit alongside identity providers, data loss prevention systems and application stacks, but the tools that gate them still assume a human is deciding what happens next. Eve's pitch is that pre-flight policies cannot fully anticipate what an agent will do once it starts chaining tools and calling APIs, and that observability without enforcement leaves security teams watching incidents in flight rather than stopping them.

How the platform works

Eve's runtime layer sits alongside AI agent runtimes and observes the traffic between agents, data sources and downstream systems. Session tainting adapts what an agent can do next based on what it has already touched, and more than 85% of policy-matched requests are evaluated deterministically at the enforcement layer rather than deferred to a large language model. The company says it currently supports Databricks, Microsoft Copilot Studio, Amazon AgentCore, Amazon Bedrock and Glean.

Eve Security AI agent runtime platform overview

The rogue-agent problem

"You cannot secure that world simply by deciding in advance what an agent should and shouldn't do," Eve co-founder and chief executive Nadav Cornberg said in a statement. "You have to understand what it is doing, why it is doing it and have the ability to intervene while it is happening." Run Ventures partner PT Ungvichian added that runtime governance for AI agents is likely to spawn "multiple important companies" as enterprises push agents into production.

The extension follows a wave of AI security funding this year, from Fortaegis's $50M Series A to Exein's $270M Series C and adjacent moves such as Analog Devices' bid for Alif Semiconductor.

Reporting based on coverage from SiliconANGLE and PR Newswire.

Category: Cyber Security

Tags: funding Cybersecurity artificial intelligence Seed Funding AI Agents AI Security

Related Articles