Cisco to Acquire Astrix Security to Secure AI Agents and Non-Human Identities

Cisco has announced its intent to acquire Astrix Security, a company focused on securing non-human identities such as API keys, service accounts, OAuth tokens, and other credentials used by applications, services, and AI agents. The acquisition will strengthen Cisco’s ability to help customers discover, govern and secure AI agents as they become more active inside enterprise environments.

Astrix was founded five years ago and has built its platform around the identity and credential layer that connects modern systems. Cisco said credentials are now being used by AI agents to access data and execute work at scale. With AI agents starting to behave less like passive tools and more like active participants in enterprise workflows, they can now access systems, make decisions and take action, which makes identity, access and permission controls more important.

Cisco plans to use Astrix to expand its work around agentic AI security and zero trust. Astrix’s capabilities include discovery and governance for AI agents, lifecycle management for agentic access, threat detection and response for compromised credentials or out-of-scope agent behavior, and centralized secrets management across vaults and cloud environments. Cisco also says that it plans to integrate Astrix into Cisco Identity Intelligence and extend the capabilities into Cisco Secure Access and Duo Identity and Access Management.

The acquisition also connects to Cisco’s broader security portfolio. Cisco pointed to AI Defense, Secure Access, Duo Agentic Identity, firewall AI traffic inspection, Splunk automation and its recent acquisition of Galileo as part of its effort to give customers more visibility into AI activity. The company said Astrix will add more context across identity, network, application and infrastructure layers, with agent activity feeding into Splunk or other SIEM platforms for investigation and response.

For security partners, the deal is another signal that non-human identity is moving into the managed security conversation. Customers adopting AI agents will need help understanding which agents exist, what systems they can access, whether permissions are too broad and how to detect risky behavior. That could open service opportunities around AI agent inventory, credential hygiene, identity governance, zero trust policy management and ongoing monitoring for agentic activity.