CrowdStrike & Nvidia deepen AI-driven security ops pact

CrowdStrike and Nvidia have expanded their collaboration on managed detection and response (MDR), focusing on using AI agents in security operations and tools for building custom security agents.

The companies have been testing Nvidia's Nemotron models inside CrowdStrike's Falcon Complete Next-Gen MDR service. They cited early internal results of up to five times faster investigations and more than three times higher triage accuracy, based on what they described as high-confidence benign classification.

The work is part of a broader push by both firms into agentic AI-software agents that can take actions across systems and interact with other agents. Security teams and technology vendors argue these designs change the risk profile of AI deployments as agents gain access to data, APIs and operational infrastructure.

Agentic MDR

MDR services typically combine software telemetry and analytics with human analysts who investigate alerts and guide response actions. CrowdStrike positions "Agentic MDR" as a model in which AI agents handle more of the repetitive front-end analysis, while analysts retain oversight and make final decisions.

The expanded effort uses the Nvidia Agent Toolkit and open Nvidia Nemotron models. It also includes Nvidia NeMo Data Designer, which the companies described as a way to generate synthetic data from expert inputs and telemetry patterns. The goal is more consistent model behaviour across investigative workflows.

CrowdStrike will also expand Charlotte AI AgentWorks with support for Nvidia Nemotron 3 Super. Charlotte AI AgentWorks is CrowdStrike's framework for building and deploying AI agents tied to its security platform. Customers will be able to build custom agents using the newly supported model.

Testing claims

The performance claims are based on internal testing of Falcon Complete Next-Gen MDR using Nvidia Nemotron Nano and Nemotron Super. The companies said benchmarking showed up to five times faster investigations and more than three times higher triage accuracy, reducing manual triage work.

They also reported results from fine-tuning Nemotron Nano for use with Falcon LogScale, CrowdStrike's log management and analytics product. According to the companies, the fine-tuned model reached 96% accuracy when generating investigation queries in LogScale-positioned as a step toward a natural-language interface for investigation workflows.

Security operations centres often handle large volumes of alerts across endpoint, identity, cloud and network systems. Many detections require human review to determine whether activity is malicious or benign. Vendors have increasingly pushed automation and AI to address analyst shortages and rising incident-response costs, while buyers have raised concerns about reliability, accountability and the risks of automating flawed decisions.

Wider partnership

The two companies are also working on security for autonomous agents, including integrating Nvidia OpenShell, an open-source runtime, with the Falcon platform. They linked the work to a "Secure-by-Design AI Blueprint for AI Agents," presented as a set of design patterns for building agents with security controls.

The announcements were made at Nvidia GTC, the company's annual developer and partner conference. Nvidia has been expanding its enterprise AI portfolio beyond GPUs into software toolkits, model families and deployment tools for customers building applications on its platform.

CrowdStrike has been adding generative AI features to the Falcon platform under its Charlotte AI branding. It is among several security vendors pitching agentic approaches as a next step beyond copilots, which typically focus on chat-driven assistance rather than autonomous action.

Daniel Bernard, CrowdStrike's chief business officer, linked the announcement to how criminal groups and state-backed attackers are adopting AI."Adversaries are already using AI to move faster and scale their operations. The future of managed defense isn't adding more analysts - it's embedding AI agents directly into SOC operations to give analysts superpowers. With Falcon Complete Next-Gen MDR, we're applying advanced reasoning models to automate investigation and triage while maintaining expert oversight. Together with NVIDIA, we're accelerating the shift toward Agentic MDR."

Justin Boitano, vice president of enterprise AI products at Nvidia, said the work combines reasoning models and synthetic data approaches with security-operations use cases."AI reasoning models and synthetic data are transforming how enterprises operationalize intelligence. Together with CrowdStrike, we're bringing secure, autonomous AI agents into security operations - enabling organizations to reason through threats, act in real time, and continuously strengthen their cyber resilience."

Kroll, which provides incident response and cyber-risk services, also endorsed the direction of the collaboration."CrowdStrike's use of advanced AI reasoning in Falcon Complete Next-Gen MDR marks a significant step forward in managed defense. By accelerating investigations and sharpening triage accuracy, it enables our teams to deliver faster, high-quality outcomes for clients around the world."

CrowdStrike said it evaluated Nemotron models across Tier 1 analysis and investigative workflow orchestration use cases, while keeping analysts in the loop for decision-making and response.