Northstar Security Launches AI Agents With $7M Seed Round

Cybersecurity startup Northstar Security has launched out of stealth mode backed by $7 million in seed funding (SecurityWeek). The investment round was led by Wing Venture Capital, alongside participation from MVP, Outliers, and cybersecurity veterans Gerhard Eschelbeck, Renaud Deraison, and Justin Somaini.

The startup develops specialized AI agents designed to assist enterprise security operations. Rather than operating as an isolated silo, the software aggregates data across a company's existing IT and security infrastructure to create what it terms a Security Context Graph.

Updated continuously in real time, this graph maps dependencies and relationships between user identities, physical and cloud assets, software applications, known vulnerabilities, and underlying business systems. The primary goal of the graph is to highlight toxic combinations of security risks that appear minor when analyzed separately.

These toxic scenarios occur when separate gaps—such as an unmanaged device or an account holding excessive administrative rights—combine to expose sensitive operational software. For example, the software can identify an unmanaged laptop assigned to a user who maintains permission to access critical financial databases.

Analysts query Northstar's autonomous agents using natural language to conduct investigations. Following an investigation, the agents can carry out automated or human-supervised actions, such as deploying endpoint security software, notifying affected employees, or revoking account privileges.

The system also exposes its context graph to third-party AI agents through the Model Context Protocol (MCP), enabling teams to build customized security workflows. Northstar offers native integrations with more than 70 enterprise IT and security environments, including AWS, Google Cloud, GitHub, Workday, Okta, Microsoft Defender, and Entra.

Northstar was co-founded by chief executive officer Amer Deeba and chief technology officer Ed Rowe. Deeba previously co-founded Normalyze and served as a senior executive at Qualys, while Rowe served as CTO of Farmers Business Network and led engineering at AppDynamics. According to Deeba, "Security teams are drowning in data but starving for context". The company will direct the new funds toward hiring, platform engineering, and go-to-market initiatives.

Northstar Security's integration of the Model Context Protocol suggests that AI-driven security architecture is rapidly shifting toward open context sharing rather than proprietary software silos. By allowing external agents to query its context graph via MCP, Northstar appears to be positioning its data model as an interoperable foundation for automated SecOps workflows. This approach likely reflects a broader recognition among security leaders that context integration, rather than generating additional isolated alerts, will determine the utility of enterprise AI security agents.

Spotted an error? See our corrections policy or email editorial@lordofclaude.com.