Agentic software control startup Neo Security Inc. said today it’s exiting a years-long spell in “stealth” mode after raising $100 million in funding.
The round was led by top tier venture capitalists in Andreessen Horowitz and Bessemer Venture Partners, and saw participation from Craft Ventures and Merlin Ventures. They’re betting that enterprises are going to need purpose-built security to safeguard their increasingly autonomous artificial intelligence systems, and likely need it soon.
Built by former executives and engineers from SentinelOne Inc., Wiz Inc. and Palo Alto Networks Inc., Neo aims to provide security operations teams with the inventory, posture intelligence, attribution and policy controls needed to manage and secure AI agents across entire organizations.
The growing adoption of autonomous AI agents that can conduct work unsupervised on behalf of humans is happening at such rapid speed that traditional security systems cannot keep pace, and it means that the attack surfaces of organizations are changing and expanding faster than ever. Enterprises are adding AI-native tools, but the bigger concern is that even the traditional software they use is integrating agentic capabilities.
Because of this, SecOps teams are being asked to secure autonomous software that’s operating inside approved applications, but they lack the necessary control and visibility to do that properly. As per Gartner Inc., just 5% of enterprise applications had agentic features at the end of last year, but more than 40% will by the end of 2026.
Co-founder and Chief Executive Nicholas Warner said existing enterprise security systems were built for a world where software behaved in a predictable way, but that’s no longer true when those tools have agentic features within them. “AI agents and agentic capabilities are being embedded into browsers, developer tools, SaaS platforms and traditional applications, giving software the ability to reason, act, invoke tools and move through workflows with valid user permissions,” he said.
That means that traditional software is altogether much riskier than before. It’s this challenge that Neo wants to take on, introducing a real-time control layer that can fend off attacks on these evolving tools. It works by giving SecOps teams a way to inventory the full suite of agentic tools running across their enterprise, even as they add new agent functionality. That way, they can understand their updated capabilities and the new risks they introduce, assess whether they’re still configured safely, and then change their security policies in the event that a tool is no longer deemed safe.
