NVIDIA Showcases Local AI Hardware Lineup at GTC Berlin
At GTC Berlin, NVIDIA is showcasing its full lineup of local AI hardware designed to power personal agents, developer applications, and enterprise workloads across Europe.

Bringing Local AI Hardware to Europe
At GTC Berlin, NVIDIA is showcasing its comprehensive lineup of local AI hardware designed to accelerate coding, engineering, robotics, and AI factory development. The hardware ecosystem spans across personal PCs, developer workstations, and enterprise-grade infrastructure. Attendees and developers can explore various sessions at GTC Berlin to learn more about secure, local systems.
The company's hardware rollout encompasses three core tiers: RTX Spark for personal agents on Windows PCs, DGX Spark for developers building and scaling applications on Linux, and DGX Station for enterprise teams running frontier workloads. Additional information and European announcements can be found by reviewing all NVIDIA in Europe announcements.
RTX Spark Brings Personal Agents to Windows PCs
The RTX Spark platform, which launched last week, introduces a new beginning for the PC by running personal agents, accelerated content creation, and high-performance gaming with up to 128GB of unified memory and 1 petaflop of AI compute. Early reviewers have praised the machine as a dream machine for serious creators and a system designed for the future of AI.
To support users and enthusiasts building local agents, the platform is backed by official playbooks. Developers can access RTX Spark playbooks to guide their implementation of local agent workflows and customized applications.
Furthermore, desktop applications like Underdog are bringing fully local, private AI experiences to the hardware. Powered by its Husky model-specific inference engine, the desktop agent runs models entirely on device so personal agents can execute tasks without user data leaving the PC. Users can download the Underdog application directly to get started.
DGX Spark Expands Developer Capabilities
For developers building and scaling applications on Linux, DGX Spark provides robust local compute. New DGX Spark 64GB systems are rolling out from partners starting Friday, October 23, allowing users to run advanced models such as Nemotron 3.5 Bolt 74B 24GB, Qwen3.8 27B, and Gemma 4 26B-A4B.
When workloads require greater capacity, developers can connect multiple DGX Spark units via built-in networking and cluster assistants. This capability enables small teams to share a single cluster backend, delivering substantial unified memory to run frontier-class models.
DGX Station Delivers Enterprise Intelligence
For larger teams and enterprise environments, DGX Station offers massive coherent memory and heavy AI performance, enabling developers to run advanced models and serve numerous users simultaneously. The system integrates closely with enterprise security and management tools to help organizations streamline deployment.
Additionally, a Windows-based version of the enterprise workstation is arriving in December, bringing local AI into the applications and operating environments that businesses already rely upon. This upcoming release extends support to various manufacturing, simulation, and engineering tools.
Sources
- NVIDIA BlogBringing the Spark to Europe: Local Intelligence Powers Next Wave of Agents
Continue chronologically





