The new AI kill switch enables security teams to intervene when an AI agent deviates from its intended behaviour #
CrowdStrike (NASDAQ: CRWD) shares climbed after the cybersecurity firm unveiled a new AI 'kill switch' designed to shut down rogue AI agents, block them from restarting, and cut off their access to corporate networks.
The launch, alongside the planned acquisitions of two AI startups for a reported $495 million, marks an aggressive push to secure businesses as they increasingly leverage autonomous AI systems for sensitive operations.
The new AI kill switch is part of CrowdStrike's Falcon Guardian platform. It gives authorised security teams the ability to intervene when an AI agent strays from its intended behaviour, rather than relying solely on alerts and investigations after something goes wrong.
'AI safety is how you get the full benefit of AI. When you deploy agents with a kill switch, you stay in control,' CrowdStrike CEO George Kurtz had said.
How CrowdStrike's AI Kill Switch Works #
AI agents can operate across corporate systems, access applications and data, and execute sequences of tasks with limited human intervention. However, an agent that behaves unexpectedly or exceeds its authorised instructions could create serious security problems, even without an external cyberattack.
CrowdStrike's new feature is designed to give security teams a direct way to contain that risk.
Through Falcon Guardian, an authorised administrator can terminate a running agent at its gateway, prevent it from launching again on a targeted host or group of hosts, and revoke its access to the enterprise network. Each intervention is recorded in an audit trail, allowing security teams to track the response.
The broader objective is to give businesses greater visibility into what AI agents are instructed to do, what they actually execute, and which corporate resources they access.
CrowdStrike Bets $495 Million on AI Security #
The AI kill switch product launch was accompanied by agreements to acquire two AI security startups, Fiddler AI and Dreadnode.
According to The Wall Street Journal, CrowdStrike is paying approximately $400 million for Fiddler AI and $95 million for Dreadnode, bringing the combined transaction value to around $495 million. The company did not disclose the individual purchase prices, saying only that both deals would be funded predominantly in cash, with a portion paid in stock subject to vesting conditions
Fiddler AI develops technology that helps businesses monitor and understand the behaviour of AI systems. Its capabilities are expected to strengthen CrowdStrike's ability to track AI agents, examine their decisions, and identify potentially problematic activity after deployment.
Elsewhere, Dreadnode specialises in testing AI models, applications, and agents for security vulnerabilities. Its technology is expected to help CrowdStrike identify weaknesses before AI systems are deployed.
AI Adoption Creates a New Cybersecurity Market #
CrowdStrike's expansion comes as businesses increasingly experiment with agentic AI, which can independently perform tasks and interact with corporate systems.
However, improperly configured or compromised agents could access sensitive information, execute unauthorised actions, or trigger a chain of activity across connected enterprise systems.
For cybersecurity companies, this creates an opportunity to sell new protections as organisations move beyond conventional chatbots towards AI systems capable of taking action. CrowdStrike is positioning Falcon Guardian to address that demand, combining visibility, security policies, and the ability to stop potentially dangerous agent activity. In all, the message behind its latest launch is straightforward: businesses can embrace autonomous AI without surrendering control over what those systems are allowed to do.
© Copyright IBTimes 2026. All rights reserved.