Obsidian Raises $85 Million Series D to Scale AI Agent Security Growth

Every enterprise is racing to put AI agents to work, not just to answer questions, to act, to log into Salesforce, pull from Databricks, update Workday, close the loop in Slack. Agents are unlocked when they can access and operate inside these third party applications. That's where AI actually pays off. Agentic access to third party applications is also where legacy platforms fall short. And this is where Obsidian shines. Combining deep understanding of identity privilege activity inside third party applications with data from agentic platforms like OpenAI, Anthropic, Cursor, Google Gemini, AWS Bedrock, we can now provide guardrails and runtime enforcement to allow enterprises to unlock agentic productivity without risking catastrophic scenarios. And this is not a niche problem. Across many organizations, more than 75% percent of enterprise applications are now third party applications. Nonhuman identity outweighs human identity by 144:1. Agentic access to third party applications is a core fundamental problem. Third party applications aren't just traditional SaaS anymore. They're data platforms running AI workloads, AI native systems like Anthropic and OpenAI, AI native apps like Harvey, Sierra, and Eleven Labs. Every team across enterprise is racing to adopt them and agents are fast becoming their primary users. And it's only accelerating. The models behind these agents aren't just getting sharper. Every new generation reasons more, plans more, can chain things together faster than any human ever could. This is not a small jump. That is a different category of capability and it's moving into production faster than any technology we have seen. The agents will have autonomy, will have access and permission to update and delete, and your data in third party applications will be the most vulnerable. The opportunity of agentic AI is clear and the guardrails are not. You get that right and agents become the biggest productivity unlock your business has ever had. And that's Obsidian Security. We believe innovation should not cost security. That's why we are on a mission to make AI adoption unstoppable by governing access, identity, and data across every third party application that runs your business. Companies like T-Mobile, Workday, S&P Global, and many of the Fortune 500 and Global 2000 customers already trust us to accelerate their AI initiatives by securing data in third party applications, which is why today we are announcing an $85 million in Series D funding led by Crescent Cove Advisors and participation from all our existing investors. This is a significant milestone for us. And none of this happens without the people who got us here. To our customers and partners, thank you for trusting us. Before Agentic Security even had a name, you told us what you needed to move fast with confidence, and we built it. To the Obsidian team, you have shipped through a market moving faster than almost any we have seen without losing sight of the discipline that got us here. And to all our existing investors, Greylock, Menlo Ventures, Norwest, IVP, Wing Ventures, and Google Ventures, thank you for the conviction to back us. You're not just funding our growth, you're funding the security infrastructure, the agentic era runs on. So where are we headed next? Every enterprise is about to hand agents more autonomy than it has ever handed any piece of software, and that will need a system they can trust. That's what Obsidian is built to be. The system enterprises trust to govern, enforce runtime controls on AI agents acting inside the third party applications safely, precisely, and without slowing the business down. The agentic era is here. Let's unlock it securely.

‍

Company valued at $1.1B with world’s largest and most complex enterprises deploying agents with Obsidian’s AI security platform
Trusted by 60 of the Fortune 500 companies including major financial institutions, social media networks and top telecom providers to prevent rogue agent activity in third-party applications
Series D funding, led by Crescent Cove Advisors, fuels Obsidian's R&D to secure enterprises against next-gen frontier models exploiting vulnerabilities in third-party applications

‍

PALO ALTO, Calif. — August 4, 2026 — Obsidian Security, the leading platform securing non-human identities and AI agents across third-party applications, today announced $85 million in Series D financing led by Crescent Cove Advisors, with participation from all existing investors including Greylock Partners and Menlo Ventures. The round follows accelerating customer growth with more than 100 customers now spending over $100K annually and over 14 customers spending more than $1 million today. As enterprises adopt AI agents built on next-gen models across platforms like Anthropic's Claude, OpenAI’s ChatGPT, and Microsoft Copilot Studio, they are increasingly turning to Obsidian to secure how those agents operate inside the third-party applications that run their business.

‍

Obsidian has spent years preparing for this moment. Non-human identities outnumber human identities 144 to 1 inside third-party applications, and agents built on frontier models are multiplying that gap further, landing in an environment where fixing exposure already takes months. As enterprises adopt agents from multiple AI ecosystems at once, agent misbehavior is becoming more common, and security teams need one durable control point across all of them. That need has driven Obsidian's growth among the world's largest organizations across the Fortune 500 and Global 2000 and that scale is now a strength in itself. As more enterprises adopt Obsidian, its network intelligence compounds, turning patterns spotted at one company into faster protection for every customer on the platform.

Obsidian will use the funding to extend that lead, bringing its platform to more of the Fortune 500 and Global 2000 as the standard for securing what AI agents can access and do.

‍

“AI is fundamentally changing how enterprises operate, creating new infrastructure challenges that legacy security models weren’t designed to address,” said Jun Hong Heng, Founder and Chief Investment Officer of Crescent Cove Advisors. “Obsidian recognized that shift early and is building the platform enterprises need to securely adopt AI at scale. The company embodies the kind of foundational technology we seek to back, and we are excited to support the team as they continue to define the future of AI security.”

‍

"AI agents gravitate toward third-party applications. That's where the data lives, that’s where the work happens, and that's exactly where the risk lives too," said Hasan Imam, CEO of Obsidian Security. “Frontier and open-source models alike are pushing agents deeper into these environments faster than most security teams can track. More than 70% of our customers already let agents into third-party apps, and that number is only going up. We intend to be the platform that protects enterprises from agents going rogue in third-party applications, so enterprises get the full return on every agent they deploy. That’s the future we’re building toward.”

‍

Closing the Gap Between AI Agent Adoption and AI Agent Governance

‍

Enterprise AI agents, whether built on platforms like Microsoft Copilot Studio, Salesforce Agentforce and n8n, or operating as autonomous developer agents like Anthropic’s Claude Code and Cowork, increasingly reach into data warehouses (Databricks, Snowflake), developer infrastructure (GitHub, GitLab), CRM, sales systems (Salesforce, HubSpot), and collaboration tools (Notion, Slack) and other third-party applications. These applications hold source code, regulated data, and intellectual property and an unsecured agent can leak, modify, or delete that data at machine speed.

‍

Security teams consistently report three top concerns: agents taking destructive or irreversible actions, unsanctioned agents connecting to critical systems without approval, and agents reaching sensitive data they were never meant to access. Recent incidents underscore the stakes from an AI agent that triggered a 13-hour cloud outage after being granted overly broad permissions, to an agent that deleted decades of irreplaceable personal files, to an agent that circumvented an application's native guardrails to delete a company's production database and backups.

‍

Obsidian has been building toward exactly this moment. Its runtime governance detects and blocks privilege escalation, excessive data access and policy violations for agents built on agentic platforms like Microsoft Copilot, n8n, Google Vertex, Amazon Bedrock, OpenAI and others. That’s backed by complete inventory of every agent, MCP server and LLM running in third party systems, so security teams know what’s connected before something goes wrong. Today’s announcement builds directly on that foundation.

Obsidian’s New AI Security Capabilities

‍

1. Agent Access Governance to Claude Code and Cowork: Building on existing coverage for platforms like Copilot Studio, OpenAI, Salesforce Agentforce, n8n, and more, Obsidian extends its AI agent security to Anthropic's Claude platform, giving security teams the ability to discover, govern, and enforce what Claude Code and Cowork agents can access and do inside their most critical third-party applications. The capability set includes restricting dangerous agent permissions to production data, managing agent access to sensitive files, right-sizing excessive access, controlling unsanctioned MCP and tool usage, and preventing destructive agent actions at runtime.

‍

2. Runtime Protection for AI Agents: Enforce real-time guardrails for agents built with Microsoft Copilot, and Anthropic Claude, detecting and blocking privilege escalation, excessive data access, and policy violations at execution time, before impact occurs, based on risk factors aligned to OWASP standards and industry best practices. Security cannot credibly govern agents if control only happens after misuse. With real-time enforcement from Obsidian, governance shifts from reactive monitoring to preventative action.

‍

3. MCP Server Inventory: Gain complete visibility into every MCP server across the enterprise, mapped to the agents that invoke them, enabling security teams to identify unsanctioned MCP usage and assess the downstream blast radius of agentic connections to backend systems and services. By surfacing the execution layer behind agents, Obsidian extends governance to the infrastructure and tools agents actually invoke, exposing all agentic risk without stitching together fragmented dashboards.

‍

4. LLM Inventory: Track every large language model powering agents in the environment to spot when models are switched or substituted, giving security and compliance teams continuous assurance that only sanctioned models drive enterprise agents. Security teams will no longer be surprised by new agentic behavior. They will always know which models are operating inside their environment.

Additional Resources

‍

Watch Obsidian Security CEO, Hasan Imam share more about today’s announcement at: https://youtu.be/jEQZwaa8coA

‍

About Obsidian Security

Obsidian Security secures the AI-first enterprise, governing data, AI and agents inside third-party applications. Trusted by Fortune 1000 and Global 2000 enterprises, Obsidian gives security teams visibility into what's connected, controls to enforce what's allowed, and the runtime context to govern AI and agent activity before it causes damage inside critical business systems. Headquartered in Palo Alto, California, Obsidian is backed by Crescent Cove Advisors, Menlo Ventures, Norwest Venture Partners, IVP, Greylock, GV, and Wing. For more information, visit www.obsidiansecurity.com.

‍