Obsidian Security hits unicorn status at

Obsidian Security raised $85 million in a Series D led by Crescent Cove Advisors, with existing investors Greylock Partners and Menlo Ventures also participating, pushing the company past a $1.1 billion valuation and total funding beyond $200 million. Obsidian says it now has more than 100 customers paying north of $100,000 annually for its platform, with 14 of those accounts topping $1 million a year, and that its customer base includes 60 companies from the Fortune 500 list.

Alongside the raise, Obsidian is extending its governance features to cover Anthropic’s Claude Code and Cowork specifically, so security teams can limit which production data and files a coding agent is allowed near and stop unapproved MCP or tool usage the moment it happens instead of catching it in a post-incident review.

What it means for operators

Obsidian built its business on SaaS security and is now pointing that same governance model at coding agents specifically, which matters if your team has adopted Claude Code or Cowork for real work without a permissions layer between the agent and your production systems. This is the second AI-agent-security unicorn minted this week, after Zenity’s $125M round, and it’s a concrete signal that “what can this agent actually access” is becoming a standard enterprise procurement question rather than something teams figure out after an incident, the same lesson from the Hugging Face agentic breach that’s still working its way through the industry.