[Pre-Order] LEADTEK AI Supercomputer 126X6000000 NVIDIA Grace Blackwell 128GB 4TB

LEADTEK NVIDIA Grace Blackwell 20-Core 128GB 4TB

NVIDIA DGX™ Spark is a new class of computer engineered from the ground up to build and run artificial intelligence (AI) workloads. Powered by the NVIDIA GB10 Grace Blackwell Superchip, DGX Spark delivers up to 1,000 AI TOPS for AI workloads.

With 128 GB of coherent unified system memory, developers can experiment, fine-tune, or perform inference with the latest large-scale generative AI models from DeepSeek, Meta, Google, and other developers, with up to 200 billion parameters (200B parameters). Additionally, NVIDIA ConnectX™ networking can connect two DGX Spark AI computers together to support inference on models up to 405 billion parameters (405B parameters).

To provide a familiar experience for developers, DGX Spark uses the same software architecture as industrial-grade AI factories, running on NVIDIA DGX OS, which uses Ubuntu Linux, and comes pre-installed with NVIDIA's latest AI software stack, with access to developer programs such as NVIDIA NIM and NVIDIA Blueprints.

Developers can immediately begin working with common tools such as PyTorch, Jupyter, and Ollama to prototype, fine-tune, and perform inference on DGX Spark, and seamlessly deploy to DGX Cloud or any accelerated data center and cloud infrastructure.

Features

NVIDIA GPU, CPU, Networking, and AI Software Technologies

Prototyping
Develop, test, and validate AI models and applications

Powered by the NVIDIA AI software stack, NVIDIA DGX Spark is a robust platform for developers to build, test, and validate AI models and AI-enhanced applications. When ready for final refinement or production deployment, systems can be seamlessly scaled to NVIDIA DGX Cloud or other data center and cloud environments utilizing NVIDIA acceleration technology.

Fine-Tuning
Fine-tune AI models up to 70 billion parameters

Elevate the performance of pretrained models by fine-tuning them on NVIDIA DGX Spark. With 128 GB of unified system memory, you can fine-tune models up to 70 billion parameters, enabling the creation of AI solutions optimized for your specific use cases.

Inference
Test, validate, and perform inference with large AI models up to 200 billion parameters

The 5th generation Tensor Core, supporting FP4 precision, delivers up to 1,000 TOPS of AI processing power. Combined with 128 GB of system memory, it accelerates inference of the latest leading-edge AI models, allowing for rapid testing, validation, and deployment directly from NVIDIA DGX Spark.

Data Science
Accelerate end-to-end data science workflows with NVIDIA RAPIDS™

From data preparation and model training to inference and deployment, NVIDIA RAPIDS on NVIDIA DGX Spark makes the development process faster and more efficient. With zero-code-change acceleration and familiar APIs, it's easy to speed up existing workloads, build customized solutions, and seamlessly scale to data centers or the cloud.

Edge Applications
Develop Edge Applications with NVIDIA AI frameworks such as Isaac™, Metropolis, and many more

NVIDIA DGX Spark is an ideal platform for building solutions in robotics, smart cities, and computer vision. By supporting NVIDIA frameworks such as Isaac, Metropolis, and Holoscan, developers can harness the full potential of DGX Spark to rapidly create and deploy edge applications.