We built NVIDIA Sentry on BlueField-4 to secure AI agents from a layer they can’t reach or tamper with. Agents can drift from their designed intent, leak data, or escape their sandbox, making independent runtime protection essential. DOCA Tower inspects agent communications, and DOCA Argus monitors the agent’s execution environment from inside the silicon. Learn how we built an independent security layer for AI agents at runtime ➡️ nvda.ws/4jzSUBV

📣 @googlecloud announces general availability of NVIDIA Vera Rubin infrastructure. Coming next month, A5X instances bring together NVIDIA Vera Rubin NVL72, NVIDIA ConnectX-9 SuperNICs and the Google Cloud Virgo Network for training, inference and large-scale reinforcement learning. Discover how @IneffableLabs is advancing “superlearner” systems and how @WPP is bringing next-gen creative agents to A5X using NVIDIA Nemotron and NVIDIA OpenShell. ➡️ nvda.ws/3VYBtRJ

Europe is building its next wave of AI infrastructure, with NVIDIA Vera CPUs at the core. At #NVIDIAGTC Berlin, @nebiusai, @nscale, and @verdacloud shared how they’re using Vera to power agents, data processing and scientific discovery, joined by a growing lineup of server makers, data platforms and developer platforms across the region. Learn more: nvda.ws/3WiFG2D

Scale AI securely across global infrastructure while maintaining full control of your models and data. At #NVIDIAGTC Berlin, we’re introducing Confidential Bare Metal with NVIDIA Confidential Computing, enabling model builders to run sensitive AI workloads on dedicated third-party infrastructure securely. @googlecloud is extending Confidential Bare Metal support to A4X Max instances, and @cohere plans to integrate it into Cohere Model Vault. ➡️ nvda.ws/4ixVneZ

NVIDIA Cloud Partners across Europe, the Nordics, Africa, and Central Asia are building full-stack AI factories that are sovereign, open and safe for AI agents. Our partners are expanding infrastructure to advance sovereign AI, scale inference, and bring physical AI into production using technologies spanning NVIDIA DSX, Vera Rubin, OpenShell, BlueField, and Spectrum-X. Read the #NVIDIAGTC Berlin announcement: nvda.ws/4z3o7Cn

👀 @ChipsandCheese9 just published a deep dive into the system architecture behind NVIDIA Vera CPU and the takeaway speaks for itself: exceptionally strong per-core performance, backed by tremendous memory bandwidth.

Join us at IBM TechXchange 2026 to see how we're collaborating with @IBM to implement enterprise AI across infrastructure, software, and services. Explore our sessions about building data platforms, deploying agents with AI factories, GPU-accelerated analytics, high-performance KV cache, and more. Learn more and register: nvda.ws/4yZiZ1Q

The @BSC_CNS’s AI Factory Accelerator Programme is now in full operation, giving startups access to sovereign compute, technical support and training across industries. Starting today, select BSC AI Factory services are also available to NVIDIA Inception members. See how startups in Spain are advancing regional innovation: nvda.ws/484FYg7

This is what packing more sandboxes per host while maintain low latency looks like. Thank you to the @modal team for sharing their benchmarks. More NVIDIA Vera CPU surprises coming!