NVIDIA has introduced DOCA Argus, a new runtime cybersecurity framework designed to safeguard AI factories—environments where complex, agentic Artificial Intelligence workflows are executed at scale. As part of the broader NVIDIA cybersecurity Artificial Intelligence platform, DOCA Argus operates on the NVIDIA BlueField networking platform to provide immediate threat detection and rapid response to attacks targeting AI workloads. The platform is built for seamless enterprise integration, providing instant threat insights and continuous monitoring without burdening existing compute resources.
DOCA Argus distinguishes itself by employing advanced memory forensics to monitor enterprise threats in real time, offering detection speeds up to 1,000 times faster than existing agentless solutions. Its agentless architecture requires no host installation or integration, enhancing resilience by operating outside the main computing environment and remaining invisible to attackers—even during potential system compromises. This zero-overhead model suits dynamic environments, including containerized and multi-tenant infrastructures, and ensures consistent security across all nodes in an AI factory.
The solution easily integrates with enterprise security frameworks such as SIEM, SOAR, and XDR, allowing cybersecurity teams to extend automation and visibility across all AI infrastructure. NVIDIA’s collaboration with Cisco further advances these goals, embedding security into the architecture of enterprise AI deployments. According to Cisco´s Jeetu Patel, the partnership aims to enable enterprises to confidently scale AI with robust safety and security mechanisms built in from inception. Complementing BlueField’s acceleration and DOCA Argus’ proactive protection is the NVIDIA Morpheus cybersecurity AI framework, together forming a full-stack, accelerated platform for enterprise defense.
With enterprises now processing vast amounts of data and deploying agentic Artificial Intelligence at scale, the need for actionable threat intelligence is critical. DOCA Argus leverages threat insights from NVIDIA’s own security team to filter out false positives, reduce alert fatigue, and focus only on validated risks. The framework is fine-tuned to protect not just traditional workloads but also containerized AI applications like NVIDIA NIM microservices. Overall, NVIDIA’s new solution aims to empower cyber defenders with real-time detection and response tools, providing a decisive security edge in the era of advanced enterprise Artificial Intelligence.
