AI BriefingNVIDIA Unveils RTX Spark
·2026.06.02 11:03
Key point
NVIDIA has unveiled the Arm-based RTX Spark Superchip, optimized for running AI agents and local LLMs.
Details
NVIDIA has announced the RTX Spark Superchip, combining a Blackwell RTX GPU with an Arm-based Grace CPU, to enter the consumer PC market.
The chip is set to be featured in Windows laptops and small desktops this fall, supporting creation, AI development, and gaming simultaneously.
Key Features and Specs:
- Hardware Configuration: Integrates up to a 6,144-core Blackwell RTX GPU and a 20-core Grace CPU into a single chip.
- AI Performance: With up to 128GB of unified memory, it can run 12-billion-parameter (12B) LLMs locally, and is optimized for running AI agents that operate in the background.
- Development Environment: The CUDA stack runs natively, enabling model development, prototyping, fine-tuning, and inference all on the same device.
- Efficiency: Designed as the most power-efficient RTX chip to date, enabling ultra-slim designs and all-day battery life.
Market and Ecosystem:
- It competes with Apple's Apple Silicon and processors from Intel and AMD, and will be released through major manufacturers including ASUS, Dell, HP, Lenovo, Microsoft, and MSI.
- It aims for a "teammate PC" concept, where the device carries out tasks once the user sets a goal.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Stanford and NVIDIA Release 'Contrastive Language Model' CLM with 9x Faster Inference