Eve Security, a runtime security company focused on identifying and stopping dangerous AI agent behavior in real time, announced that it has raised $4.5 million in new funding, extending its seed round to $7.5 million in total funding. The new financing was led by Run Ventures, with participation from Dreamit Ventures and Blu Ventures, along with continued investment from LiveOak Ventures.
Eve Security is developing security infrastructure designed to help enterprises understand, govern, and control autonomous AI agents while those agents are operating. The company said demand is increasing as businesses connect AI agents to sensitive corporate data, applications, and infrastructure, creating new security risks that traditional cybersecurity products were not built to address.
The funding follows a year in which Eve focused on validating its approach to runtime AI agent security and developing its technology alongside enterprise customers and security leaders. The company said interest from both customers and investors has increased as autonomous AI systems become more capable and gain broader access to enterprise environments.
Unlike conventional security tools that primarily monitor users, identities, endpoints, applications, and infrastructure, Eve is designed to analyze the behavior of autonomous agents. Traditional tools may detect a compromised credential or suspicious command, but determining whether a sequence of otherwise legitimate actions by an AI agent represents normal execution or emerging dangerous behavior presents a different challenge.
Eve provides enterprises with governance, observability, and runtime controls for AI agents. Its platform can identify high-risk or anomalous activity, investigate why an agent is performing particular actions, and intervene before an action reaches a critical corporate system.
The company’s investment thesis was further validated during the fundraising process when Run Ventures introduced Eve to a group of chief information security officers. Their feedback helped reinforce the company’s view that enterprises will require a dedicated runtime security layer as agent deployments become increasingly autonomous.
Eve began its formal go-to-market efforts in January 2026, and the company said existing customers are already expanding their use of the platform. Eve views these expansions as early evidence that runtime AI agent security could become a persistent enterprise requirement rather than a narrowly focused security product.
The company is also expanding its platform to address increasingly complex agent deployments. New capabilities include session tainting, which dynamically adapts and restricts what an agent can do based on its exposure to sensitive information and its previous actions.
Eve has also expanded discovery, enforcement, and automated remediation capabilities across platforms including Databricks, Glean, Microsoft Copilot Studio, Amazon AgentCore, and Amazon Bedrock.
Another core component of Eve’s platform converts enterprise security policies into a deterministic enforcement layer. The company said more than 85% of policy-matched requests can be evaluated and enforced deterministically, helping enterprises apply consistent controls without relying entirely on probabilistic AI decision-making.
For situations requiring additional analysis, Eve can incorporate real-time information from identity providers, data loss prevention systems, Databricks, Snowflake, and other enterprise data platforms. This approach combines deterministic controls with contextual information to determine whether an agent should be permitted to complete a particular action.
The new funding will primarily support go-to-market expansion and revenue growth. Over the next 12 to 18 months, Eve plans to demonstrate across multiple enterprise customers that its runtime AI security technology can produce repeatable and measurable results.
KEY QUOTES:
“Agents are becoming extraordinarily capable, and enterprises are giving them access to increasingly valuable systems. The OpenAI incident showed what can happen when an agent’s pursuit of an objective takes it somewhere its creators never intended. You cannot secure that world simply by deciding in advance what an agent should and shouldn’t do. You have to understand what it is doing, why it is doing it and have the ability to intervene while it is happening.”
Nadav Cornberg, Co-Founder and CEO of Eve Security
“AI runtime security is developing into a significant new security category, and we believe there will be multiple important companies built in this market. Eve recognized early that securing autonomous systems requires understanding and controlling behavior at runtime, and the company is exceptionally well positioned to capitalize on that shift.”
PT Ungvichian, Run Ventures
“We had high conviction in the technical thesis. What we wanted next was market proof. We’re now seeing that pull directly from CISOs and customers. This round lets us put significantly more resources behind turning that early demand into a repeatable business.”
Nadav Cornberg, Co-Founder and CEO of Eve Security
