CrowdStrike Buys Fiddler AI and Dreadnode for $495 Million to Secure AI Agents

CrowdStrike Fiddler Dreadnode AI kill switch https://www.pexels.com/photo/red-button-on-the-chrome-finished-trim-9661246/

Industry leader CrowdStrike has recently taken several major steps to secure the agentic AI space. On November 3rd, 2026, at the company’s Fal.Con cybersecurity conference in Barcelona, CrowdStrike announced three moves: the acquisitions of Fiddler AI and Dreadnode, and the AI Kill Switch. The terms of either deal have not been disclosed; the Wall Street Journal reported a $495 million total, with one source citing $400 million for Fiddler and $95 million for Dreadnode, while others cite unnamed sources reporting $300 million and $200 million price tags, respectively. The deals are being carried out mostly in cash, with some stock subject to vesting.

Agents With Valid Credentials and Their Own Agenda

Certain problems arising in the threat landscape have led CrowdStrike to take these steps to protect agentic AI usage. Many agents have been observed ignoring instructions, breaking out of sandboxes, and claiming to have successfully completed work that they never finished. Auditors, regulators, and boards are increasingly asking to be able to verify what actions agents have actually carried out.

These factors have created a landscape where agentic AI is at a major risk of acting outside its scope, without much recourse for visibility and accountability regarding unauthorized activity. The identity of the actor carrying out an action proves access, not intent. The industry has demonstrated a critical need for improved management and security of AI agents.

Fiddler Serves as the “Flight Recorder”

Fiddler AI was founded in 2018 and has raised around $99 million since then. When integrated into CrowdStrike’s portfolio, the tool will be rebranded and sold as Falcon Observe. It is expected to be available to customers within the same quarter that both acquisitions are finalized, which is set to occur in CrowdStrike’s fiscal Q4 2027.

George Kurtz, CEO and founder of CrowdStrike, referred to Fiddler AI as “the flight recorder for every agent.” It enables teams to trace agentic reasoning paths, tool calls, behavioral drift, and token usage. This bolsters agent visibility, monitoring, accountability, and governance, allowing for more secure agentic AI deployment and management.

Dreadnode Tests Models Before Attackers Break Them

Dreadnode, founded in 2023, has raised $14 million in a Series A funding round in 2025. Led by CEO Brad Palm, Dreadnode’s team will be joining CrowdStrike’s Cyber Superintelligence Lab, with the platform’s capabilities expected to also become available in CrowdStrike’s Q4 2027, which ends on January 31st, 2027.

The startup develops offensive security tools to test AI models against a range of harmful activity. Users define what they would like tested, and Dreadnode runs over 100 attack strategies, including jailbreaks and prompt injection. The findings of these tests are applied to several leading threat information repositories like the OWASP Top 10 for LLMs and MITRE ATLAS. Dreadnode returns these results, ranked and mapped onto these existing frameworks, as policy recommendations.

What the Kill Switch Does and Doesn't Cover

Announced on the same day as the two acquisitions, CrowdStrike’s “kill switch” for AI agents is an extra layer of assurance for many organizations with AI agents in production. It offers gateway controls, relaunch blocking, and network cutoff capabilities to allow security teams more control over agentic AI operations.

The kill switch is treated as a significant responsibility—it requires an authorized administrator to trigger it, and each administrator action is written to an audit log to ensure ongoing accountability. The human trigger and the dependence on gateways to protect against agentic threats are factors that create gaps in the ability of the kill switch to mitigate a breadth of AI agent-related risks.

Platform Assembly Around AI Security

The developments announced at Fal.Con recently are all in line with CrowdStrike’s strategy of the past few years to consolidate a range of capabilities under its portfolio for the ongoing growth of AI tool and agent usage. The company acquired AI security leader Pangea in September 2025, for a reported price of $260 million, followed by the acquisition of SGNL, which closed in February 2026 to strengthen identity security.

CrowdStrike’s strategy reflects recent moves made by the company’s rivals on the market to buy their way into agent security. Similarly to CrowdStrike, other industry leaders like Palo Alto Networks have also been making a spate of deals in recent years to acquire tools and platforms built for agentic AI security, governance, and management.

Questions CISOs Should Ask Now

In the months leading up to the closing of CrowdStrike’s recent acquisition deals, security leaders should be watching the space and ensuring that their own policies align with oncoming AI agent security developments. Agent inventory and observability should be established before any purchase is made. Many questions remain in the air regarding the integration timelines of the two acquisitions and the demonstrable capabilities of agents already in production. Enterprises are also advised to get information on the packaging and pricing of the integrated products, such as how much the acquisitions will be integrated into a consolidated platform versus remaining independently controlled.

Author
  • PJ Bradley
    Contributing Writer, Security Buzz
    PJ Bradley is a writer from southeast Michigan with a Bachelor's degree in history from Oakland University. She has a background in school-age care and experience tutoring college history students.