CrowdStrike Acquires Fiddler AI and Dreadnode to Build Out AI Agent Security Platform
CrowdStrike announced agreements to acquire Fiddler AI and Dreadnode, two AI security companies, in deals revealed at its Fal.Con Europe conference in Barcelona. Terms were not disclosed by CrowdStrike, though the Wall Street Journal reported pricing of roughly $400 million for Fiddler and $95 million for Dreadnode. The acquisitions target an emerging gap in AI agent security: agents can hold valid credentials and approved tools yet still make poor decisions, hallucinate completed work, or ignore instructions, a problem traditional code-focused security testing does not address. Fiddler, which will become Falcon Observe, provides post- and pre-deployment observability, tracing agent decisions, flagging drifting or fabricated behavior, and generating audit records, with support for frameworks like LangGraph, Amazon Bedrock, and Google ADK. Dreadnode provides pre-deployment adversarial testing, running over 100 known attack techniques including jailbreaks and prompt injection against models and agents, mapped to standards such as the OWASP Top 10 for LLMs and the NIST AI Risk Management Framework.
Detalhes técnicos
Alongside the acquisitions, CrowdStrike introduced new Falcon Guardian runtime protections for AI agents and an AI Kill Switch feature, described by CEO George Kurtz as a control existing outside the model at runtime that allows humans to halt any AI agent regardless of the agent's own behavior or attempts to resist. Kurtz framed this as a shift from traditional security (keeping adversaries out) to a broader AI safety mission (allowing agents to operate at speed under human control).
Remediação
Both Fiddler and Dreadnode products are expected to become generally available within CrowdStrike's fourth fiscal quarter of 2027, with the acquisitions expected to close in the same period pending standard regulatory approvals. Payment will be made mostly in cash, with a stock component vesting over time based on performance conditions.
Linha do tempo
-
CrowdStrike Acquires Fiddler AI and Dreadnode to Build Out AI Agent Security Platform
ler a análise desta fonte ocultar
CrowdStrike announced agreements to acquire Fiddler AI and Dreadnode, two AI security companies, in deals revealed at its Fal.Con Europe conference in Barcelona. Terms were not disclosed by CrowdStrike, though the Wall Street Journal reported pricing of roughly $400 million for Fiddler and $95 million for Dreadnode. The acquisitions target an emerging gap in AI agent security: agents can hold valid credentials and approved tools yet still make poor decisions, hallucinate completed work, or ignore instructions, a problem traditional code-focused security testing does not address. Fiddler, which will become Falcon Observe, provides post- and pre-deployment observability, tracing agent decisions, flagging drifting or fabricated behavior, and generating audit records, with support for frameworks like LangGraph, Amazon Bedrock, and Google ADK. Dreadnode provides pre-deployment adversarial testing, running over 100 known attack techniques including jailbreaks and prompt injection against models and agents, mapped to standards such as the OWASP Top 10 for LLMs and the NIST AI Risk Management Framework.
Technical Details
Alongside the acquisitions, CrowdStrike introduced new Falcon Guardian runtime protections for AI agents and an AI Kill Switch feature, described by CEO George Kurtz as a control existing outside the model at runtime that allows humans to halt any AI agent regardless of the agent's own behavior or attempts to resist. Kurtz framed this as a shift from traditional security (keeping adversaries out) to a broader AI safety mission (allowing agents to operate at speed under human control).
Remediation
Both Fiddler and Dreadnode products are expected to become generally available within CrowdStrike's fourth fiscal quarter of 2027, with the acquisitions expected to close in the same period pending standard regulatory approvals. Payment will be made mostly in cash, with a stock component vesting over time based on performance conditions.