CrowdStrike and Nvidia unveil Secure-by-Design AI Blueprint to lock down autonomous AI agents

crowdstrike-and-nvidia-unveil-secure-by-design-ai-blueprint-to-lock-down-autonomous-ai-agents
CrowdStrike and Nvidia unveil Secure-by-Design AI Blueprint to lock down autonomous AI agents
CrowdStrike mobile

CrowdStrike has introduced a Secure-by-Design AI Blueprint created with Nvidia that integrates security protections directly into Nvidia’s OpenShell runtime for AI agents. It combines the CrowdStrike Falcon platform with Nvidia’s infrastructure so that organizations can run autonomous AI agents with monitoring and controls, regardless of whether those agents run locally or in the cloud.

The architecture connects Falcon security capabilities with Nvidia OpenShell, an open-source runtime environment that manages policy enforcement for AI agents. This runtime creates isolated sandboxes for agents and allows organizations to apply guardrails around how autonomous systems interact with data, applications, and infrastructure.

SEE ALSO: CrowdStrike report shows ransomware surging across Europe

Interest in AI agents is growing rapidly as companies move beyond “copilots” and toward systems that can reason and take actions automatically. Those capabilities also lead to new security concerns because agents often operate with privileged access to resources such as internal data, APIs, and computing systems. Traditional security models were built around far slower workflows and fixed access controls, and they don’t always fit with systems that can make decisions and act continuously.

Secure-by-Design AI Blueprint enforcement

The Secure-by-Design AI Blueprint embeds monitoring and enforcement across the AI stack so that protections operate during development as well as during runtime. The idea is to give organizations visibility into how agents behave and make sure that all activity stays within defined policies as tasks are executed.

Integration between Falcon and OpenShell means security controls can operate directly inside the runtime environment where AI agents function. OpenShell provides isolated environments and private inference capabilities, while Falcon extends monitoring and protection to the agents themselves.

Agents running locally on Nvidia DGX Spark or DGX Station systems can be monitored and protected through Falcon endpoint security. The same security architecture can also extend to cloud deployments built using Nvidia’s AI-Q Blueprint, which is intended for large-scale AI research and production workloads.

The blueprint includes several layers of controls to track activity and limit unsafe behavior. These include AI Detection and Response capabilities that monitor prompts, responses, and agent actions in real time, as well as endpoint protections that watch system activity on machines running local AI agents.

Cloud environments receive additional monitoring via Falcon cloud security tools, which provide runtime controls across AI infrastructure and workloads. Identity-based controls are also part of the blueprint, with Falcon identity security managing how agents authenticate and interact with services, APIs, and datasets.

Daniel Bernard, Chief Business Officer at CrowdStrike, said the move toward autonomous agents requires a different approach to security. “As we enter the agentic era, agents no longer simply assist — they act,” he said. “This shift fundamentally changes the security equation, and security must be embedded into the AI stack itself. Together with NVIDIA, we are delivering a Secure-by-Design architecture that enables organizations to operationalize agents with confidence and control.”

Justin Boitano, Vice President of Enterprise Platforms at Nvidia, described the technology as part of a wide push to make AI agents safer as adoption grows. “Autonomous agents will fundamentally reshape how we work,” he said. “By integrating CrowdStrike’s security platform with the NVIDIA Agent Toolkit, we’re enabling enterprises to build and scale safer, autonomous AI agents to help transform their operations, empower every employee, and securely generate intelligence at the speed of business.”

The two companies are also exploring intent-aware controls that evaluate how agents plan and carry out tasks, limiting the potential impact of unintended actions while still allowing agents to operate independently within defined boundaries.

James Higgins, Chief Information Security Officer at CoreWeave, said large-scale AI infrastructure requires clear governance as adoption increases. “AI infrastructure is moving from experimentation to mission-critical production,” he said. “As we scale GPU-accelerated environments, AI agents must be observable, governed, and resilient by design. The collaboration between CrowdStrike and NVIDIA secures AI systems at the foundation — enabling high-performance AI environments without compromising control.”

What do you think about CrowdStrike’s Secure-by-Design AI Blueprint? Let us know in the comments.

Image credit: rafapress/depositphotos.com