Xage Security announced support for new and enhanced NVIDIA DOCA security capabilities and NVIDIA Vera BlueField-4 STX, also announced. Xage’s integration with NVIDIA DOCA security is designed to enable line-speed visibility, policy enforcement, and control across AI factory environments, supercharging Xage’s Zero Trust for Agentic AI to observe, govern, and control AI interactions at massive scale. The combined solution is designed to deliver these outcomes while preserving AI performance and operational efficiency, without requiring modification of host-based software or the protected workloads.
NVIDIA Vera BlueField-4 STX, powered by NVIDIA DOCA security, defines a new class of secure-by-design AI storage, embedding real-time protection for data, agents and context memory in silicon to enable trusted agentic AI.
“AI factories are becoming the new critical infrastructure—and they require critical-infrastructure-grade security protection,” said Duncan Greatwood, CEO of Xage Security. “As autonomous AI agents gain access to sensitive data, APIs, applications, and core systems, organizations need unbypassable visibility into and control over what those agents can see, do, and change. By supporting NVIDIA Vera BlueField-4 STX and Enhanced DOCA security architecture, Xage brings identity-based Zero Trust to the heart of the AI Factory, delivering the foundational security required for the successful use of AI with organizations’ mission critical data, workflows and operations.”
Xage Zero Trust for Agentic AI helps enterprises move AI agents safely from pilots to production. Agentic AI systems increasingly operate across APIs, SaaS platforms, databases, cloud services, file storage, internal applications, and operational technology environments. Xage capabilities provide end-to-end visibility and control across the full AI interaction chain: users, agents, models, tools, APIs, applications, infrastructure, and resources. Unlike approaches focused only on prompts or model outputs, Xage governs exactly what each agent can access and what actions it can take.
NVIDIA Vera BlueField-4 STX, powered by NVIDIA DOCA security, provides a foundation for applying granular security policies at line rate, detecting and blocking threats to AI workloads in real time, and implementing visibility and control to secure agentic workflows at scale.
Xage’s Zero Trust for Agentic AI will provide multiple integration points with NVIDIA BlueField-4 STX, enabling end-to-end protection for AI factories. The company will deliver AI Security Gateways that run natively within the NVIDIA DOCA architecture through the Xage Resource Gateway, Xage Agent Sentry, and Xage Extended Protection (XEP) components, securing both resources and AI agents.
The Xage Resource Gateway can also integrate with NVIDIA DOCA Vault to provide file access visibility and control, while the Xage Fabric policy engine leverages information from DOCA to make real-time interaction decisions.
In addition, Xage can integrate with and provide identity-level intelligence to DOCA Argus and DOCA Flow, sharing security-relevant events such as login attempts, entitlement delegation, and credential changes to enhance threat detection capabilities on NVIDIA Vera BlueField-4 STX.
For hardware-accelerated enforcement, Xage integrates with DOCA OvS, dynamically configuring the BlueField OvS stack to permit authorized interactions while blocking unauthorized activity. Xage can also serve as the response mechanism for real-time automated detection and response, enabling actions such as isolating suspicious AI agents, suspending access for potentially compromised users, or applying additional protections to resources exhibiting anomalous access patterns.
Together, Xage and NVIDIA DOCA integration will create a closed-loop security model: DOCA-based infrastructure observing runtime behavior for visibility; Xage evaluating identity, policy, resource, context, and action; and Xage enforcement controls, working through DOCA-OvS, helping to allow, deny, contain, or escalate agent activity before it creates business risk.
Xage can help organizations using NVIDIA Vera BlueField-4 STX and NVIDIA DOCA monitor, control, and protect AI agents at both the infrastructure and workload levels. The platform enforces least-privilege access across AI systems, resources, data, tools, APIs, and applications, while applying granular ‘just-in-time, just-enough’ zero trust policies at line rate to block threats.
It also enables governance of interactions between agents, models, APIs, storage systems, data, and other resources, helping to prevent unauthorized activity. By blocking lateral movement and containing suspicious agents, Xage helps limit the spread of potential threats. The platform gathers detailed operational and security information to support anomaly detection, while enabling organizations to implement and enforce automated threat response measures. Together, these capabilities help maximize the efficiency, security, and resilience of AI factories.
Xage Zero Trust for Agentic AI gives enterprises a way to secure AI agents, govern access to critical resources, prevent unauthorized actions, and maintain audit-ready visibility across the AI lifecycle. With support for NVIDIA DOCA and NVIDIA Vera BlueField-4 STX, Xage extends its distributed Zero Trust architecture into the next generation of AI Factories—helping enterprises, government agencies, and critical infrastructure operators adopt agentic AI with greater confidence, resilience, visibility, and control.
