Onyx Security Lands $113M to Govern Enterprise AI Agents

A Big Bet on Securing Autonomous AI

Onyx Security has raised $113 million in a Series B funding round aimed at helping enterprises govern and secure AI agents, bringing the startup's total funding to $153 million 1. The scale of the raise signals how urgently investors and enterprise buyers are treating the challenge of controlling autonomous software that can act, decide, and execute tasks with far less human oversight than traditional applications. As companies race to deploy AI agents across customer service, coding, and internal operations, the question of who is watching those agents — and how — has become a pressing governance problem rather than a theoretical one.

Why Control, Not Just Capability, Is the New Priority

The funding lands at a moment when enterprise security teams are realizing that AI adoption cannot be separated from infrastructure hardening. One recurring theme in current security commentary is that AI systems are only as trustworthy as the underlying technical plumbing that connects them to enterprise data and services. Analysts point specifically to APIs as the soft underbelly of AI deployments, arguing that unless organizations lock down API access and authentication rigorously, every downstream AI security investment — including tools like Onyx's — remains vulnerable 2. In other words, agent-control platforms are necessary but not sufficient; they sit atop a broader security stack that many enterprises have yet to fully secure.

Adoption Is Uneven, and That's a Governance Risk Too

Controlling AI agents assumes that AI is being adopted in a structured, visible way — but that assumption is increasingly shaky. Reporting on enterprise AI usage patterns shows a widening access gap, with senior leaders enjoying far greater access to AI tools and training than junior employees, producing uneven adoption and strategic misalignment within organizations 3. This disparity matters directly to the security conversation: when adoption is fragmented and inconsistent, shadow use of AI agents becomes more likely, making centralized oversight tools like Onyx's more valuable but also harder to implement comprehensively across an organization.

Trust Deficits Beyond the Tech Sector

The caution around AI autonomy is not confined to corporate IT departments. In healthcare, adoption has notably lagged despite AI's promise for diagnostics, personalized treatment, and drug development, largely because of deep-seated trust concerns tied to patient safety and accountability 4. Related coverage outlines specific missteps — from opaque decision-making to poor communication with clinicians — that have eroded confidence in healthcare AI systems, even as the underlying technology matures 5. These parallel struggles across industries underscore a common thread: whether in hospitals or corporate back offices, the barrier to AI adoption is increasingly about governance, transparency, and trust rather than raw technical capability.

The Bigger Picture

Taken together, the funding news and the surrounding commentary suggest enterprise AI is entering a phase where control mechanisms, secure infrastructure, equitable access, and institutional trust are becoming as important as the AI models themselves. Onyx Security's raise reflects investor confidence that demand for agent governance will only grow, but the broader coverage makes clear that no single tool can solve adoption challenges rooted in uneven access, insecure APIs, and lingering skepticism about handing consequential decisions to autonomous systems.