CrowdStrike to Acquire Dreadnode to Bring Automated AI Security Testing to the Falcon Platform

Today, at Fal.Con EU 2026, I’m excited to share CrowdStrike’s intent to acquire Dreadnode. This acquisition will bring automated AI security testing to the CrowdStrike Falcon® platform so our customers can validate the security of AI systems before they go into production.  

Modern businesses deploy AI models and agents before confirming they’re secure, an approach that can have severe consequences. When an agent ships with a weakness that allows jailbreaking or manipulation, it provides the adversary with an opportunity to reach the sensitive systems and data the agent can access. These weaknesses must be discovered before the agent goes into production. Security testing built for conventional software does not find them. 

Dreadnode does. Its technology tests AI models, applications, and agents with known attack techniques to learn where they break so fixes can be applied before deployment. CrowdStrike, together with Dreadnode, will extend the Falcon platform into the layer where businesses build AI so every customer can evaluate their AI agents for security gaps and deploy them with confidence.

Dreadnode will be generally available in Q4 FY27. Let’s take a closer look at what it will bring.

Automated AI Security Testing on the Falcon Platform

The probabilistic and dynamic nature of AI agents and LLMs has broken the traditional security testing model, which has long been used to assess code that consistently behaves in the same way. This testing approach can’t identify weaknesses inside the reasoning loop built into AI models and agents.

AI has also introduced new security risks: Prompt injection, jailbreaking, and agent tool manipulation can lead to customer data loss, financial loss, and brand reputation damage. Dreadnode was built to understand the risks to AI systems by probing for vulnerabilities to uncover these issues before adversaries can.

How Dreadnode Works 

Users first describe what they want to test in plain language. Dreadnode runs more than 50 attack techniques — including jailbreaks, prompt injection, and multi-turn and agent-specific attacks — against the AI systems, models, model repositories, applications, agents, and MCP servers the business relies on before they reach production.

With the testing complete, Dreadnode ranks findings by severity and maps them to the OWASP Top 10 for LLMs, OWASP Agentic Security (ASI01-ASI10), MITRE ATLAS™, and NIST AI RMF. All assessments are aggregated into a single risk picture that includes total findings, attack success rates, severity breakdown, finding outcomes, and deep risk metrics. Dreadnode also delivers an exportable report so stakeholders can make go/no-go decisions.

Key to this process is human-in-the-loop review. Model builders can expand any findings to see the exact attacker prompt and target response, and walk through attacks step-by-step. Operators can reclassify their findings, adjust attack severity levels, and document reasoning. All dashboard metrics will recompute automatically when findings are reclassified.

These findings will fortify CrowdStrike Falcon® Guardian, our flagship AI detection and response (AIDR) solution, which protects enterprise-developed AI at runtime. Dreadnode results will feed into Falcon Guardian as policy recommendations, so a weakness found in testing informs runtime protection. Dreadnode testing data will also strengthen CrowdStrike’s SafeMind models, advancing our research into securing AI systems.

Stronger Security for Enterprise-Developed AI 

Businesses are integrating AI models, agents, and applications into daily operations faster than security teams can validate them. Traditional software security testing isn’t built to capture the risks these technologies can present. With Dreadnode, CrowdStrike will extend the Falcon platform to give enterprises a fast, repeatable way to test their AI systems before they’re deployed.

CrowdStrike’s mission has always been to stop breaches. As new technologies emerge and the threat landscape evolves, we must adapt to provide our customers with the robust defense they need. Today, that means delivering solutions that secure AI end-to-end, from development to deployment to runtime. Last year, we acquired Pangea to secure enterprise AI use and development. Last month, we announced Falcon Guardian to deliver runtime protection for AI agents, introduce a new AI gateway, and extend our elite security services to AI systems.

The acquisition of Dreadnode and its automated AI security testing will be a critical step on our journey to protect businesses in the AI era. We are eager to welcome the Dreadnode team. 

Forward-Looking Statements

This blog contains forward-looking statements, including statements regarding the closing and benefits of the proposed acquisition. These statements involve risks and uncertainties, and actual results may differ materially. There are a number of risks that could cause actual results to differ materially, including the satisfaction of the acquisition’s closing conditions, our ability to integrate Dreadnode, and other risks described in the filings we make with the U.S. Securities and Exchange Commission from time to time.