Pluto Security exits stealth with $42M to police employee-built AI tools
Your biggest AI security risk is not hackers. It is your own employees building AI apps in secret. Pluto just raised $42M to watch them.
Funding · Source: PYMNTS
What happened
Pluto Security just exited stealth with $42 million in total funding. Jump Capital led the recent Series A round of roughly $34 million. TLV Partners led the seed round. The startup built an endpoint security platform specifically for the agentic era to secure the AI tools your employees are building.
The traction is massive for a company founded in 2025. Pluto hit multi-million dollar annual recurring revenue within six months of commercial operations. They already secure over 1.5 million endpoints worldwide. They signed more than 40 enterprise customers, including Oscar Health, Workato, Sixt, and several Fortune 500 companies.
The founders know enterprise security inside out. CEO Shahar Bahat and her co-founders Gil Maman and Tal Ofer Regev come from Microsoft, Palo Alto Networks, and Israel's elite Unit 8200. Their platform analyzes AI context in real time. It blocks dangerous actions without disabling the actual AI tools employees need to do their jobs.
Key facts
- $42 million — Total funding raised by Pluto Security
- 1.5 million — Endpoints currently secured by Pluto worldwide
- 15 — Average number of AI tools used per employee
- 40+ — Enterprise customers signed in less than a year
Why it matters
The way we build software has fundamentally changed. Employees in marketing and sales are now acting like developers. They use tools like Claude Code, Lovable, Codex, and Base44 to spin up custom workflows and agents. The average employee now uses 15 different AI tools. This creates a massive shadow AI supply chain. Security teams are flying blind because they cannot vet every new model, extension, or Model Context Protocol component fast enough.
The old security playbook was binary. You either blocked a tool entirely or let it run wild. Pluto proves the market demands a nuanced approach. Security must now understand the context of an AI agent's actions. If an agent tries to move sensitive data or access authentication credentials, the system blocks the action but keeps the application functional. This shift means enterprise adoption of AI will accelerate because IT finally has a safety net to manage the chaos.
For builders
Build context-aware security tools
The binary block-or-allow security model is dead. Enterprises will pay massive premiums for tools that understand what an AI agent is actually trying to do. If you build security products, focus on real-time action filtering so security teams can stop bad actions without killing productivity.
Target non-developer AI workflows
Marketing and sales teams are building their own apps. They do not know how to secure them. Founders who build guardrails specifically for these non-technical builders will find eager buyers in enterprise IT departments who are desperate to regain control.
Secure the shadow AI supply chain
Employees are plugging random extensions and Model Context Protocol components into corporate systems. This is a massive vulnerability. Startups that can audit and secure these third-party AI dependencies will win big enterprise contracts.
My take
I keep saying that your own team is your biggest vulnerability. You cannot stop marketing and sales from building AI apps, and if you try, your company will lose its competitive edge. Pluto gets it exactly right by policing the action, not the tool.
Original reporting: PYMNTS. This is my rewrite and opinion.