As enterprises more and more undertake AI, securing AI factories — the place complicated, agentic workflows are executed — has by no means been extra essential.
NVIDIA is bringing runtime cybersecurity to each AI manufacturing facility with a brand new NVIDIA DOCA software program framework, a part of the NVIDIA cybersecurity AI platform. Working on the NVIDIA BlueField networking platform, NVIDIA DOCA Argus operates on each node to right away detect and reply to assaults on AI workloads, integrating seamlessly with enterprise safety programs to ship prompt risk insights.
The DOCA Argus framework supplies runtime risk detection by utilizing superior reminiscence forensics to watch threats in actual time, delivering detection speeds as much as 1,000x sooner than current agentless options — with out impacting system efficiency.
In contrast to standard instruments, Argus runs independently of the host, requiring no brokers, integration or reliance on host-based sources. This agentless, zero-overhead design enhances system effectivity and ensures resilient safety in any AI compute setting, together with containerized and multi-tenant infrastructures. By working outdoors the host, Argus stays invisible to attackers — even within the occasion of a system compromise.
Cybersecurity professionals can seamlessly combine the framework with their SIEM, SOAR and XDR safety platforms, enabling steady monitoring and automatic risk mitigation and lengthening their current cybersecurity capabilities for AI infrastructure.
NVIDIA BlueField is a foundational safety element for each AI manufacturing facility, offering built-in, data-centric safety for AI workloads at scale. By combining BlueField’s acceleration capabilities with DOCA Argus’ proactive risk detection, enterprises can safe AI factories with out compromising efficiency or effectivity.
Cisco is collaborating with NVIDIA to ship a Safe AI Manufacturing unit with NVIDIA structure that simplifies how enterprises deploy and shield AI infrastructure at scale. The structure embeds safety into each layer of the AI manufacturing facility, guaranteeing runtime safety is inbuilt from the beginning somewhat than bolted on after deployment.
“Now’s the time for enterprises to be driving ahead with AI, however the important thing to unlocking revolutionary use circumstances and enabling broad adoption is security and safety,” mentioned Jeetu Patel, govt vice chairman and chief product officer at Cisco. “NVIDIA and Cisco are offering enterprises with the infrastructure they should confidently scale AI whereas safeguarding their most precious information.”
DOCA Argus and BlueField are a part of the NVIDIA cybersecurity AI platform — a full-stack, accelerated computing platform purpose-built for AI-driven safety. It combines BlueField’s data-centric safety and Argus’ real-time risk detection with NVIDIA AI Enterprise software program — together with the NVIDIA Morpheus cybersecurity AI framework — to ship visibility and management throughout an AI manufacturing facility. It additionally faucets into agentic AI to autonomously understand, purpose and reply to threats in actual time.
Optimized AI Workload Risk Detection
Enterprises are inundated with large volumes of information, making it troublesome to pinpoint actual threats. The rising adoption of agentic AI, with AI fashions and autonomous brokers working at enterprise scale to seamlessly join information, functions and customers, brings unprecedented alternatives for gleaning insights from information — whereas introducing the necessity for superior safety that may preserve tempo.
DOCA Argus is fine-tuned and optimized utilizing insights from NVIDIA’s personal safety group, surfacing solely actual, validated threats. By specializing in well-known risk actors and eliminating false positives, the framework supplies enterprises with actionable intelligence, decreasing alert fatigue and streamlining safety operations.
Argus is purpose-built to guard containerized workloads like NVIDIA NIM microservices, incorporating real-world risk intelligence and validation to safe each layer of the AI utility stack.
“Cyber defenders want sturdy instruments to successfully shield AI factories, which function the muse for agentic reasoning,” mentioned David Reber, chief safety officer at NVIDIA. “The DOCA Argus framework delivers real-time safety insights to allow autonomous detection and response — equipping defenders with a knowledge benefit by way of actionable intelligence.”
Get began with DOCA Argus and meet NVIDIA on the RSA Convention in San Francisco, operating by way of Thursday, Could 1.