CrowdStrike agrees to buy Fiddler AI, which helps monitor AI agents, for ~$400M, and Dreadnode, an automated security AI testing tool, for ~$95M, mostly in cash

Your AI agent's decisions and actions can now be traced and audited, and security teams can shut them down immediately if they go rogue.

CrowdStrike has announced its intention to acquire two AI security startups, Fiddler AI for approximately $400 million and Dreadnode for approximately $95 million, primarily in cash. Fiddler AI offers capabilities for monitoring and explaining AI agent decisions and behaviors, which will be integrated into CrowdStrike's Falcon platform as Falcon Observe. Dreadnode provides automated security testing for AI models, applications, and agents, identifying vulnerabilities like prompt injection and jailbreaks, and will be incorporated into Falcon Guardian. Both acquisitions are anticipated to conclude in CrowdStrike's fiscal fourth quarter of 2027. Concurrently, CrowdStrike has launched an AI Kill Switch feature within Falcon Guardian, allowing security teams to immediately halt rogue AI agents by cutting off network access and preventing re-launches. The company has also expanded its AI security integrations with Anthropic and Amazon Web Services.

These acquisitions signal a significant move by CrowdStrike to bolster its AI security offerings, aiming to provide enterprises with greater control and visibility over the burgeoning AI agents they deploy. By integrating Fiddler's explainability and Dreadnode's testing capabilities, CrowdStrike is positioning its Falcon platform as a comprehensive solution for managing the risks associated with AI, from pre-deployment vulnerability assessment to post-deployment behavior monitoring and control.

The introduction of the AI Kill Switch, alongside these acquisitions, addresses a critical emerging need for immediate intervention against potentially harmful AI agent actions. This move is likely to influence other cybersecurity vendors to enhance their own AI oversight and control mechanisms, potentially setting a new standard for securing AI deployments in enterprise environments and offering a tangible safeguard against unexpected AI behavior.