NVIDIA launches DSX POD AI factory design and OpenShell agent security architecture
Article timeline
The articles in this story. Times are ET.
🚨 AI News | TestingCatalogXNVIDIA introduces OpenShell reference architecture and DSX POD for faster AI factoriesAINVIDIA has introduced an OpenShell Reference Architecture and a new NVIDIA DSX POD. The company says AI factory deployment at a prepared site now takes 14 days, down from six months.
NVIDIAOfficialNVIDIA introduces OpenShell Reference Architecture to set boundaries for AI agentsAINVIDIA introduces the OpenShell Reference Architecture, which helps businesses set boundaries for AI agents and stop them when they go beyond those limits. The post presents the architecture as a way to secure agents, with no further technical details given.
NVIDIA AIOfficialNVIDIA publishes OpenShell reference architecture for secure AI agentsAINVIDIA's OpenShell Reference Architecture combines compute for running AI agents at scale with runtime controls that govern their access to tools and data. The post says the hardware and software are designed to work together, with agents making fast tool calls under security controls they cannot rewrite.
- NVIDIA Technical BlogOfficialNVIDIA OpenShell reference architecture secures and governs AI agent workloads at scale
AINVIDIA's Technical Blog presents the OpenShell reference architecture for securing and governing AI agent workloads at scale. The post says long-running autonomous agents can access files, tools and the internet, and spawn other agents, which compounds risk as they download code, build tools, request privileges and reach more infrastructure.
- NVIDIA NewsroomOfficialNVIDIA launches OpenShell Reference Architecture for safe AI agents
AINVIDIA announced the OpenShell Reference Architecture, a full-stack setup for building, containing, observing, and enforcing AI agents on enterprise and edge infrastructure. It combines Vera Rubin for inference, Vera CPUs for agent harnesses, and BlueField-4 DPUs running NVIDIA Sentry for in-silicon policy enforcement. The architecture is now available, and partners including Anthropic, Microsoft, and OpenAI are working with it.
- NVIDIA NewsroomOfficialNVIDIA DSX reference designs cut AI factory cooling water and power use in Europe
AINVIDIA says its DSX AI Factory Platform, adopted by European partners, uses closed-loop liquid cooling and dry coolers to reach near-zero facility cooling water consumption. DSX MaxLPS software lets operators generate the same number of tokens with up to 30% less power, with CoreWeave, Google Cloud, Nebius and others planning to use it. At 2-gigawatt scale, the design can save 54 million liters of water a day compared with conventional cooling towers.
- NVIDIA NewsroomOfficialNVIDIA introduces DSX POD AI factories built in modular containers for faster deployment
AINVIDIA introduces DSX POD, a reference design for complete Vera Rubin AI factories that partners build and validate inside modular containers before delivery. Factory-built units cut deployment time at a prepared site from six months to as little as 14 days. NVIDIA says a global group of partners, including Bull, Dell Technologies, HPE, Lenovo, Schneider Electric and Vertiv, will deliver the modular AI factories worldwide in early 2027.
- NVIDIA BlogOfficialNVIDIA unveils DSX POD, agent safety architecture and new robotics platform at GTC Berlin
AIAt GTC Berlin, NVIDIA CEO Jensen Huang announced NVIDIA DSX POD, a reference design for prebuilt Vera Rubin AI factories that the source says can cut onsite deployment from six months to 14 days. He also introduced the NVIDIA OpenShell Reference Architecture for Safe AI Agents and Agile ONE SN, a 31-degree-of-freedom reference humanoid expected next year. The source says NVIDIA committed nearly $10 billion to Europe's AI ecosystem across its weeklong European tour.