Northstar Security Exits Stealth With $7M Seed for AI Risk Agents

Northstar Security has launched publicly with $7 million in seed capital led by Wing Venture Capital, betting that AI agents — not more dashboards — will define the next decade of cybersecurity.

Wing headed the round. MVP and Outliers also invested, alongside individual checks from cybersecurity veterans Gerhard Eschelbeck, Renaud Deraison and Justin Somaini. The company plans to spend the money on product buildout, headcount growth and scaling commercial operations.

What problem is Northstar attacking?

Enterprises have poured billions into tools that flag threats, weak points and configuration errors. The catch: that information sits in separate systems. An unpatched flaw, an unmanaged laptop or an over-privileged account might each look minor alone. Together, they can open a serious path to critical business systems.

Northstar calls these "toxic combinations," and finding them today is brutal manual work. Analysts must gather evidence from numerous tools, trace connections, judge business consequences and organize remediation — a process that can stretch for hours.

How does the platform actually work?

The core is a live Security Context Graph that charts links between assets, identities, vulnerabilities, applications and business systems. People and AI agents work from this shared picture, letting the platform answer intricate questions, surface concealed exposures and run autonomous remediation — with human sign-off where needed.

Northstar offers a concrete example: a corporate laptop lacking endpoint protection. Establishing its owner, its system access and its risk level could normally mean working through six separate tools. On Northstar, an analyst asks in plain language — and might learn the device belongs to a privileged employee with access to a critical payroll system. AI agents can then trigger fixes: installing protection, alerting the owner or limiting access.

The platform bundles three functions:

• Connected security context — unifying assets, identities, vulnerabilities, configurations and business relationships in one graph.
• Risk intelligence — ranking issues by likely business impact.
• Agentic remediation — coordinating fixes across teams and systems.

Northstar also exposes its graph to external AI agents through Model Context Protocol (MCP), so organizations can build their own workflows with Claude and other AI tools. It currently offers more than 70 integrations, including AWS, Google Cloud, Microsoft Defender, Microsoft Entra, Okta, GitHub, Slack, Workday and Zendesk.

Who's behind it?

Founders Amer Deeba and Ed Rowe first collaborated at Adobe in the late 1990s. Deeba previously co-founded Normalyze, later bought by Proofpoint, and held senior posts at Qualys. Rowe, now CTO, was formerly CTO of Farmers Business Network and senior vice president of engineering at AppDynamics.

CEO Deeba frames the company's thesis bluntly. "Security teams are drowning in data but starving for context. We've spent decades building better tools to find security problems, but connecting the dots and fixing them remains painfully manual. We believe the next generation of security won't be defined by more dashboards or alerts," he said.

"It will be powered by AI agents that understand the full context of an organization's environment and can take proactive action. That's why we built Northstar: to provide the context foundation for agentic security and move the industry from finding problems to continuously fixing them."

The seed money gives Northstar runway to prove that context-first, agent-driven remediation can cut those hours-long investigations to minutes. If it works, the pitch stops being another alert feed and becomes a real rollback of the industry's manual triage burden.