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 (James Rundle/Wall Street Journal)

CrowdStrike has agreed to acquire two AI-security companies. Fiddler AI helps monitor AI agents, while Dreadnode provides an automated tool for testing AI security. The purchases extend CrowdStrike’s capabilities beyond conventional cybersecurity protection and into systems built with artificial intelligence.

The larger deal is Fiddler AI, priced at approximately $400 million. Dreadnode will cost about $95 million. Together, the acquisitions target two different needs: watching AI agents as they operate and testing AI systems for security problems before or during deployment.

The deals come as companies deploy AI agents faster than some security teams can establish controls. That timing gives CrowdStrike a reason to build products for this expanding risk area. The article says the purchases will be made mostly in cash, although it does not provide further payment details.

The announced prices add up to approximately $495 million: about $400 million for Fiddler AI and about $95 million for Dreadnode. That combined figure is an estimate because both prices are described with approximate symbols, not as exact final amounts.

“Mostly in cash” means most of the purchase consideration will be paid directly in money. A smaller portion could involve another form of consideration, such as shares or another agreed payment structure, although the article does not specify what that portion will be. The wording describes the payment mix, not a separate discount.

The cash-heavy structure makes the headline value easier to understand: CrowdStrike is committing roughly $495 million in total, with most funded in cash. The deals are aimed at AI security as companies deploy agents faster than some security teams can create controls. The article gives no closing date or financing details.

Fiddler AI is a company that helps monitor AI agents. Monitoring means observing how these systems operate, what they do, and whether their behavior stays within a company’s intended boundaries. The source does not describe Fiddler’s specific product features, but it places the company in the growing AI-security market.

For example, a company might deploy an AI agent to carry out tasks across internal systems. A monitoring tool could help security teams see the agent’s activity and identify behavior that needs review. The key mechanism is visibility: teams cannot easily control systems they cannot observe. The article does not name a particular incident or monitoring alert.

The need is becoming more urgent because companies are deploying AI agents faster than some security teams can establish controls. CrowdStrike’s purchase suggests monitoring may become part of broader cybersecurity protection. The article does not say exactly how CrowdStrike will integrate Fiddler’s technology or what customers will receive after the deal.

Dreadnode offers an automated security AI testing tool. In practical terms, such a tool evaluates an AI system’s security without relying entirely on manual checks. The source does not list the exact threats, tests, or technical methods involved, so its precise testing scope cannot be determined from the article.

A concrete example would be testing an AI system before a company puts it into production. Automated checks could look for unsafe behavior or weaknesses in the system’s defenses, then give security teams findings to investigate. The central mechanism is repeatable testing at machine speed, rather than relying only on occasional human review. This example explains the category, not a specific Dreadnode feature stated in the source.

Dreadnode’s tool fits a market where AI deployment is moving faster than security controls. CrowdStrike agreed to buy the company for about $95 million. The acquisition could expand CrowdStrike’s ability to test AI security, but the article does not state the product roadmap or the exact risks Dreadnode detects.

When AI agents spread faster than security controls, companies can lose the ability to manage new systems safely and consistently. The article frames this as a timing problem: deployment is accelerating, while some security teams are still building the rules and safeguards needed to oversee agents.

For example, an organization could put an AI agent into service before it has reliable monitoring or testing. Security staff might then struggle to see what the agent is doing or identify weaknesses in its design. This is an illustrative example, since the source does not describe a specific breach, failure, or harmful action. The underlying mechanism is a control gap.

That gap helps explain CrowdStrike’s purchases of Fiddler AI and Dreadnode. One company helps monitor agents, and the other offers automated AI security testing. Together, those capabilities could support stronger oversight as deployments grow. However, the article does not report any resulting incident or predict a specific outcome for customers.

CrowdStrike’s purchases could connect its established cybersecurity business with tools built specifically for AI systems. The source does not explicitly state CrowdStrike’s strategy, but the timing suggests a response to companies adopting AI agents faster than some security teams can create controls. Monitoring and testing address risks that ordinary protection may not fully cover.

For example, a company could use CrowdStrike for endpoint protection while using Fiddler AI to observe an agent and Dreadnode to test its security. The mechanism would be complementary coverage: protect the computing environment, watch the agent’s behavior, and examine the AI system for weaknesses. This example describes a possible combination, not an integration announced in the article.

The acquisitions would give CrowdStrike positions in both AI-agent monitoring and automated AI security testing. That could make its offerings more relevant as businesses deploy agents. Still, the article does not state product integration plans, customer benefits, or whether the companies will remain separate after the purchases.

An AI agent is a software system that can receive a goal, decide what steps to take, and perform actions to pursue that goal. A chatbot usually focuses on a conversation and responds to prompts. A traditional program follows rules written in advance, while an agent can use an AI model to choose among steps in less predictable situations. These distinctions are general definitions, not supplied by the article.

For example, a chatbot might answer a question about an order. An AI agent could go further by checking a system, finding the order, and starting an approved follow-up action. A traditional program might perform the same workflow only through fixed rules. The agent’s action-taking ability is why monitoring and security testing become important, although the article gives no specific example.

The source focuses on companies deploying AI agents faster than some security teams can establish controls. That creates demand for tools such as Fiddler AI’s monitoring and Dreadnode’s automated security testing. The article does not define agents technically or describe their permissions, reliability, or exact operating environments.