Nvidia has launched the RTX Spark, an ARM-based chip designed to run 120-billion-parameter language models locally with a context window reaching 1 million tokens. The chip will target high-end laptops priced between $3,000 and $4,000, aimed at developers, researchers, and power users wanting to run sophisticated AI agents without cloud dependency. By adopting ARM architecture and leveraging its CUDA ecosystem, Nvidia is directly challenging Apple Silicon’s dominance in on-device AI. Industry analysts have framed this as a direct challenge to Apple, even though Nvidia has not explicitly named Apple as its primary rival.
Source: Read the original article
