Obsidian Security Funding: $85M Series D Hits $1.1B Valuation

A new unicorn in AI agent security

Obsidian Security has closed an $85 million Series D round that values the company at $1.1 billion, putting it in unicorn territory. 12 The startup announced the raise on Tuesday. Both outlets covering the deal tie the funding to the same trend: companies are deploying more AI agents, and those agents increasingly reach into sensitive corporate systems and data. 12

SecurityWeek describes Obsidian's product as a platform for governing AI agents across third-party applications. 1 Reuters, via its syndicated report, calls the company an AI security startup. It frames the round as a response to enterprises trying to secure a growing number of AI agents that access sensitive information. 2 The two accounts differ mostly in emphasis. One focuses on what the product does, and the other on the market demand behind the valuation. Together they give a consistent picture of why investors are paying a premium.

Why AI agents create a new security problem

The core issue is easy to state and hard to solve. Traditional enterprise security assumed that the main actors inside business software were people, such as employees who log in, click around, and hold specific permissions. AI agents break that assumption. They act on behalf of users, often with broad access, and they work across many software-as-a-service tools at once: CRMs, collaboration suites, code repositories, file storage, and more.

The phrase "third-party applications" in SecurityWeek's description matters here. 1 The risk is not limited to a company's own infrastructure. It sits in the sprawl of outside SaaS platforms where corporate data now lives. An agent connected to several of those services can, in principle, read, move, or change data at machine speed. If it is misconfigured, over-permissioned, or manipulated, the damage can spread before a human notices.

Reuters' point that agents are accessing sensitive data is the commercial hook. 2 Security buyers tend to spend when they can tie a tool to a concrete exposure. Agents with credentials to customer records, financial systems, or intellectual property are exactly that kind of exposure.

Reading the valuation

A $1.1 billion valuation on an $85 million raise suggests investors see Obsidian as more than a niche tool. 12 My reading is that the market is treating AI agent governance as a likely new budget line for security teams, comparable to how identity management, cloud security posture, and SaaS security each grew into their own categories.

Obsidian's positioning appears built to take advantage of that shift. By focusing on agents operating across third-party apps instead of a single vendor's ecosystem, the company is betting that enterprises will want a neutral layer of oversight. 1 The alternative would be a patchwork of controls offered by each SaaS provider separately. That argument has historically worked well in security: buyers rarely trust a platform to police itself, and they like seeing everything in one place.

Some caution is warranted. The available reporting does not detail Obsidian's revenue, customer count, or lead investors, so it is hard to judge how much of the valuation reflects current traction versus expectations about where AI adoption is heading. A Series D is a late-stage round, which implies an established business. Still, the AI framing in both reports suggests the narrative around agents is doing a lot of the work. 12

The broader context

The deal fits a pattern that has become familiar across the security industry. As generative AI moves from chat interfaces into autonomous or semi-autonomous agents that take actions, vendors and investors are racing to define what "securing AI" actually means. Some focus on model-level risks such as prompt injection or data leakage through outputs. Others, like Obsidian, focus on the access layer: which agents exist, what they are connected to, and what they are allowed to do. 1

The access-layer approach has a practical advantage. It maps onto controls security teams already understand, including permissions, monitoring, anomaly detection, and audit trails. That makes it easier to sell into existing security organizations than more abstract model-safety tooling.

The takeaway

Obsidian's raise is best read as a signal that AI agent governance is turning into a fundable, standalone security category. 12 The demand is real. Agents are proliferating inside enterprises and touching sensitive data. 2 Whether Obsidian becomes the defining vendor in this space or one of many will depend on execution, competition, and how quickly enterprises turn concern about agent risk into purchasing decisions. For now, investors are betting heavily that the concern will turn into spending.