Cisco in Talks to Acquire AI Agent Security Firm Astrix for at least $250 Mn

Cisco Systems is in advanced talks to acquire Astrix Security, a Tel Aviv–based firm focused on securing artificial intelligence agents, according to a report by The Information.

The deal is expected to be valued between $250 million and $350 million—roughly triple the total funding raised by Astrix to date. Its backers include Workday and the Anthology Fund, supported by Anthropic.

The development follows Cisco's announcement from last week about its intent to acquire AI observability startup Galileo.

Astrix develops software that helps enterprises identify and secure AI agents operating across their networks. Its platform automatically maps these agents, detects the tools they use, and identifies non-human identities such as service accounts. It also scans for vulnerabilities, including misconfigurations that could expose internal systems to the public internet.

The platform flags risky behavior, such as AI agents with excessive permissions, and presents findings through a centralized dashboard. Administrators can respond using built-in controls, including just-in-time (JIT) access policies that limit how long agents can interact with sensitive systems. For example, login credentials can be set to expire within minutes, reducing the risk of misuse.

Astrix also monitors anomalous activity, such as large-scale data extraction attempts, and can automatically revoke access for suspicious agents. Its tools integrate with third-party cybersecurity platforms to streamline incident response.

The potential acquisition follows Cisco’s recent agreement to acquire Galileo Technologies, underscoring its growing focus on securing enterprise AI systems.

Last year, Cisco acquired EzDubs, a startup that pioneered real-time speech-to-speech translation across 31 languages, to bring highly natural, AI-driven voice translation directly into its Webex collaboration suite.