Nvidia has unveiled its new RTX Spark chip, designed for personal AI agents and creative tasks, integrating a Blackwell RTX GPU with a Grace CPU via NVLink-C2C. This powerful combination offers 1 Petaflop of compute and 128GB of unified memory, enabling laptops to run 120B parameter large language models locally with a 1 million token context window. Nvidia also introduced the DGX Spark, a desktop supercomputer for AI developers, featuring similar core components but optimized for model prototyping, fine-tuning, and inference up to 200B parameters. AI
IMPACT These new Nvidia products aim to bring powerful AI capabilities, including large model inference and agent execution, directly to consumer devices and developer workstations, potentially accelerating local AI development and adoption.
RANK_REASON New hardware product release from a major AI chip manufacturer. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- 1 million token
- Bilibili World
- Blackwell RTX GPU
- DGX Spark
- Grace CPU
- Hermes Agent
- NemoClaw
- Nvidia
- NVLink – C2C
- OpenClaw
- OpenShell
- Qwen
- RTX Spark
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
