Nvidia Expands AI Supercomputer Access with Taiwan Partners for DGX Spark and DGX Station

Nvidia has announced a strategic expansion of its AI computing ecosystem by partnering with Taiwan’s top system manufacturers, including Acer, Gigabyte, and MSI, to build and distribute its DGX Spark and DGX Station AI supercomputers. These partnerships aim to broaden access to powerful, desktop-form-factor AI systems that deliver data-center-class performance, enabling a wide range of users—from developers and data scientists to enterprises and research institutions—to accelerate AI innovation.

The DGX Spark system is powered by the Nvidia GB10 Grace Blackwell Superchip and features fifth-generation Tensor Cores, delivering up to one petaflop of AI compute and 128GB of unified memory. This compact system allows seamless model exporting to Nvidia DGX Cloud or other accelerated cloud and data center infrastructures, making it ideal for pushing the boundaries of generative AI and accelerating workloads across various industries.

For more demanding AI workloads, the DGX Station features the Nvidia GB300 Grace Blackwell Ultra Desktop Superchip, offering up to 20 petaflops of AI performance and 784GB of unified system memory. It supports high-speed networking at up to 800Gb/s and can partition its GPU resources into up to seven instances, enabling flexible multi-user environments. This makes DGX Station suitable both as a powerful individual desktop and as a centralized compute node for teams.

Both DGX Spark and DGX Station run the Nvidia DGX operating system, preconfigured with the latest AI software stack, including Nvidia NIM microservices and Nvidia Blueprints. Developers can use familiar tools such as PyTorch, Jupyter, and Ollama to prototype, fine-tune, and perform inference locally before deploying models to the cloud or data centers, ensuring a seamless development workflow.

Leading global system builders like Dell Technologies and HP Inc. are among the first to develop and offer these systems, recognizing the growing enterprise demand for localized, high-performance AI computing. Dell’s Pro Max series and HP’s ZGX platform exemplify how these collaborations are redefining desktop computing by bringing data-center-class AI capabilities directly to developers and researchers.

Availability for DGX Spark begins in July through a wide range of partners including Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI, with reservations already open. DGX Station is expected to be available later this year from ASUS, Dell Technologies, Gigabyte, HP, and MSI. This expanded ecosystem ensures that enterprises, startups, and research institutions worldwide can access powerful AI computing resources tailored to their needs.

Broader Significance and Industry Impact

The introduction of DGX Spark and DGX Station marks a pivotal shift in AI infrastructure by enabling powerful, scalable AI workloads directly at the desktop level without compromising data privacy or performance. This is especially critical as agentic AI systems capable of autonomous decision-making become more prevalent, requiring robust, localized compute resources.

By bridging the gap between desktop and data center AI computing, Nvidia’s new systems empower a global ecosystem of developers and organizations to innovate faster, enhance AI model development cycles, and securely manage proprietary data. This democratization of AI supercomputing is poised to accelerate breakthroughs across industries including healthcare, finance, manufacturing, and government research.

Nvidia’s collaboration with leading hardware manufacturers and system builders also signals a maturing AI hardware ecosystem, where integration, scalability, and user experience are prioritized to meet the complex demands of modern AI workloads. This approach ensures that AI innovation is accessible, efficient, and aligned with enterprise and research needs worldwide.