Nvidia GTC Berlin: Jensen Huang Unveils DSX POD and Enterprise AI Agent Arc

Red Hat announced an expanded collaboration with Nvidia at GTC Berlin to help enterprises secure autonomous AI agents, govern model access, and deploy workloads on Nvidia's upcoming Vera Rubin data center infrastructure.

According to IT Home, NVIDIA announced at GTC 2026 in Berlin that DGX Spark now supports clustering up to eight devices with up to 1TB of unified memory. The company also partnered with llama.cpp and vLLM to boost inference performance by up to 40% on RTX PRO 6000 Blackwell and introduced the Sync Model Launcher for streamlined deployment.

NVIDIA introduced the DSX POD modular AI factory reference design at GTC Berlin. Pre-integrated with Vera Rubin systems and validated offsite, the solution cuts on-site deployment time from six months to as little as 14 days.

NVIDIA CEO Jensen Huang introduced the DSX POD modular AI factory reference design at GTC, built for the Vera Rubin platform. The solution integrates computing and cooling into modular containers in-factory, cutting deployment times to as fast as 14 days with shipments starting in early 2027.

Nvidia CEO Jensen Huang urged Europe's lagging AI industry to accelerate development and promoted a modular system designed to quickly build AI data centers for local developers, according to Bloomberg Technology.

According to ZDNET, ServiceNow and Nvidia have launched Arc, a personal and long-running enterprise AI agent designed for productivity tasks across email and Teams. Running on RTX Spark and Nvidia OpenShell, the agent incorporates multiple layers of real-time security checks.

According to IT之家, NVIDIA CEO Jensen Huang stated at GTC Berlin that software tools and companies will not be ended by AI; instead, AI agents will drive demand for more tool users. A demonstration with Dassault Systèmes showed an AI agent modifying car door designs and invoking simulation tools to test solutions.

At GTC Berlin, NVIDIA CEO Jensen Huang unveiled the Vera Rubin architecture and the NVIDIA DSX POD reference design for AI factories. Pre-integrated into modular containers by manufacturing partners, the solution reduces site deployment cycles to as fast as 14 days.

Nvidia has introduced plug-and-play Vera Rubin racks designed for standard truck shipping, along with the DSX Pod AI Factory housed in a single container. The system combines seven racks to reduce AI token service deployment time to 14 days.

According to Bloomberg, Nvidia CEO Jensen Huang urged Europe to accelerate its adoption of artificial intelligence at GTC Berlin, stating that all end markets and industries must adopt AI for the region to truly benefit from it.

As investors scrutinize the debt and economics of AI infrastructure builders, Nebius stated that AI computing prices are rising. The remarks address growing market concerns over the financial sustainability and business models of emerging neoclouds.

According to Bloomberg, Nvidia CEO Jensen Huang urged European industries to adopt artificial intelligence more quickly at the GTC conference in Berlin. Huang emphasized that creating a robust AI economy requires widespread adoption across all end markets rather than just technology supply.

According to ZDNet, ServiceNow and NVIDIA have jointly launched Arc, a new enterprise AI agent designed to tackle security challenges through multi-layered real-time safety checks.

According to East Money, NVIDIA CEO Jensen Huang showcased the company's latest developments in AI infrastructure, agents, and humanoid robotics at GTC Berlin. NVIDIA introduced the DSX POD reference design based on the Vera Rubin platform, which aims to shrink modular AI factory deployment to 14 days.

Nvidia CEO Jensen Huang urged European companies to adopt artificial intelligence more quickly. Amid growing investor skepticism over massive infrastructure spending, Huang maintained a bullish outlook, pushing European industries to embrace AI transformation.

At GTC Berlin, NVIDIA showcased its lineup of local AI hardware designed for personal and enterprise use. The announcements featured RTX Spark for personal agents on Windows PCs, DGX Spark for Linux developers, and DGX Station for enterprise teams.

According to IT Home, NVIDIA CEO Jensen Huang stated at the GTC conference in Berlin that all telecom operators will eventually provide AI compute, with base stations expected to evolve into distributed AI networks. He reiterated his five-layer AI cake theory, emphasizing that AI serves as a foundational industrial infrastructure akin to electricity and the internet.

NVIDIA has introduced the DSX POD architecture to help operators deploy high-performance compute and interconnected systems more quickly and predictably, reducing infrastructure complexity for AI workloads.

According to Bloomberg Technology, Nvidia CEO Jensen Huang delivered a speech at GTC Berlin, drawing significant industry attention to potential technological updates.