Nvidia has expanded its Enterprise artificial intelligence Factory validated design to incorporate BlueField cybersecurity and infrastructure acceleration, positioning the data processing unit as a core part of the operating system for artificial intelligence factories. The company frames this as a response to scaling artificial intelligence deployments, where infrastructure must efficiently manage data, secure each stage of the pipeline, and accelerate services that move, protect and process information in parallel with artificial intelligence workloads. The updated design is intended to keep data center services running at the same speed as artificial intelligence while bringing security up to the same performance level.
BlueField is described as the processor that accelerates the essential services behind artificial intelligence, including networking, storage, security and orchestration, so that CPUs and GPUs can stay focused on model computation. By offloading these tasks to a dedicated processor, enterprises are expected to scale artificial intelligence with consistent performance while maintaining real-time protection across the entire pipeline, from data ingestion to fine-tuning and inference. Nvidia emphasizes that BlueField delivers zero-trust, hardware-accelerated security that operates at the same speed and scale as enterprise artificial intelligence workloads, and that isolating security functions on the data processing unit strengthens defenses without disrupting operations. With the Nvidia Doca Argus framework providing artificial intelligence runtime visibility and threat detection, organizations gain continuous monitoring of runtime behavior and threats across the artificial intelligence lifecycle.
The expanded validated design integrates a roster of cybersecurity and infrastructure software partners, all validated to run on Nvidia RTX Pro Servers and built to exploit BlueField acceleration. Armis Centrix focuses on continuous cyber exposure management across artificial intelligence factories of any size, while Check Point Infinity artificial intelligence Cloud Protect provides real-time network and host security against artificial intelligence threats using Nvidia Doca Argus telemetry. F5 Big IP Next for Kubernetes targets workload isolation, policy enforcement and performance for evolving Kubernetes environments, and Fortinet FortiGate VM brings next-generation firewalling and zero-trust segmentation directly into data center infrastructure to extend Fortinet Security Fabric protections to artificial intelligence factories. Palo Alto Networks Prisma Airs aims to enforce zero-trust, artificial intelligence runtime protection at the infrastructure layer with Doca Argus, Rafay offers multi-tenant isolation and consistent controls for production artificial intelligence factories, and Red Hat OpenShift supplies a production-grade application platform with enhanced networking and security. Spectro Cloud’s PaletteAI and PaletteAI Secure focus on making full-stack artificial intelligence easier to integrate, faster to deploy and more efficient to manage, and Trend Vision One delivers infrastructure-level monitoring and policy enforcement that combines real-time telemetry with global threat intelligence. Nvidia positions this partner ecosystem, combined with BlueField acceleration, as a stronger foundation for enterprises looking to deploy secure, efficient artificial intelligence at scale.
