Nvidia stock up 4% after RTX Spark and N1X CPU reveal

Nvidia shares rose about 4% after it introduced the RTX Spark superchip, pairing a Blackwell GPU with an Arm-based N1X CPU built with Microsoft and MediaTek for on‑PC AI agents.

Nvidia unveiled the RTX Spark superchip during CEO Jensen Huang’s keynote at the Computex conference in Taiwan. The stock rose about 4% in early trading. RTX Spark pairs a Blackwell GPU with an Arm-based N1X CPU and includes 128 gigabytes of unified memory to run agentic AI on personal computers.

Nvidia described RTX Spark as “the most efficient PC chip ever built.” The N1X processor was developed with Microsoft and custom designed by MediaTek. Nvidia said the N1X will be integrated into RTX Spark and is scheduled to appear this fall in new Windows PCs from Microsoft, Dell, HP, ASUS, Lenovo and MSI.

Market reaction followed quickly: Nvidia shares climbed roughly 4% while broader indexes were under pressure. Intel shares fell about 6.5%, AMD dropped 5.2% and Qualcomm slid 9%. Arm Holdings rose about 10%.

RTX Spark combines the Blackwell GPU, used for model inference, with the Arm-based N1X CPU, intended to manage agent orchestration and local workloads. The chip’s unified memory pool is designed to allow larger models and datasets to be processed on-device rather than relying solely on cloud servers.

Huang also said Nvidia’s Vera CPU for data centers has entered full production and that the company is manufacturing millions of the chips. He listed early customers including Anthropic, OpenAI, xAI, Dell, Oracle and CoreWeave and said the data-center CPUs are expected to be available this fall. He described the processors as “going to be our new major growth driver.”

The announcements extend Nvidia’s work beyond graphics processing into broader AI computing on personal machines and in data centers. The company’s entry into PC CPUs places it in direct competition with established x86 makers and with firms expanding AI-capable PC chips, while using Arm-based architecture as an alternative to traditional PC designs.

Articles by this author