Former Palantir and Torq veterans raise $8 million Seed to secure the rise of AI agents
Arrakis Security is betting that enterprises need a new cybersecurity layer focused not just on who has access to systems, but on what autonomous AI systems actually do once they are inside.
Israeli cybersecurity startup Arrakis Security has raised $8 million in Seed funding to build a platform designed to monitor and govern the behavior of enterprise AI agents, as companies increasingly deploy autonomous systems capable of accessing sensitive data, executing workflows and making decisions without direct human supervision.
The round was led by Hetz Ventures and included investments from Mati Staniszewski, CEO of ElevenLabs; Ofer Smadari, CEO of Torq; Amitai Ratzzon, founder and CEO of Pentera; and several senior executives from Palantir Technologies.
Arrakis was founded by Tal Baron, Omer Efrat and Ron Shani, veterans of Torq and Palantir, who argue that traditional cybersecurity tools were designed around a world of human users interacting with machines, not autonomous software systems that can make decisions and act independently across enterprise environments.
The company currently employs around 20 people. The new funding will be used to expand its research, development and product teams in Israel, strengthen its detection and response capabilities, and accelerate deployments among organizations adopting AI agents.
The rise of AI agents is creating a new security challenge for enterprises. Unlike traditional software tools, these systems can independently perform tasks, access multiple systems and adapt their behavior based on changing information.
Arrakis argues that organizations are facing what it calls a “visibility illusion”: companies may know that an AI agent exists and is connected to a specific system, but often lack a detailed understanding of what the agent is actually doing, which permissions it is using, whether its actions comply with company policies and when normal behavior turns into a security risk.
The company’s platform is designed to discover AI agents operating throughout an organization, create behavioral profiles for each one, monitor their activities in real time and detect actions that deviate from expected patterns.
Rather than focusing only on securing the point where an AI agent enters an organization, such as an endpoint or application, Arrakis is targeting the broader behavior of autonomous systems across SaaS platforms, cloud infrastructure, business applications and internal workflows.
Related articles:
The platform monitors agent activity, analyzes non-human identities and compares actions against organizational policies. When it detects suspicious behavior, such as unauthorized access to sensitive information, data exfiltration, privilege abuse or unexpected service activation, it can automatically block actions, revoke permissions, suspend integrations or trigger response workflows.
“You cannot secure a machine that thinks and operates differently from humans using security tools that were designed to monitor humans interacting with machines,” said Tal Baron, co-founder and CEO of Arrakis Security.
“Organizations are rapidly deploying AI agents that access CRMs, process financial data, execute workflows and make decisions autonomously, yet in most cases they have no way to understand what those agents are actually doing.”
To support its platform, Arrakis has established Arrakis Labs, a research team composed of former Microsoft security researchers and veterans of leading technology units.
The team focuses on identifying vulnerabilities, attack techniques and behavioral patterns specific to AI agents, with those capabilities continuously integrated into the company’s platform.
