Nvidia RTX Spark SoC Leaks Show ARM Performance Beating Intel Core Ultra 9
Geekbench 7 results reveal a 20-core ARM platform that outperforms Intel's high-end x86 silicon in multi-core tests.
Nvidia is expanding beyond the GPU market to challenge the laptop silicon landscape with its new RTX Spark SoC. Recent leaks on Geekbench 7 reveal a high-performance ARM-based platform designed to bring tight CPU and GPU integration to Windows 11 laptops.
Two variants of the RTX Spark have appeared in the benchmarks: an 18-core and a 20-core ARM CPU. According to data reported by Tom's Hardware and El Chapuzas Informático, the 20-core model achieved a single-core score of 2,570 and a multi-core score of 23,126. The 18-core version followed closely, posting 2,541 in single-core and 21,776 in multi-core tests. Both tested units were equipped with 64 GB of unified memory and ran Windows 11 on ARM.
Notably, the 20-core RTX Spark demonstrated significant efficiency over traditional x86 architecture. It outperformed the Intel Core Ultra 9 285HX in multi-core performance by approximately 15.4%, scoring 23,126 against Intel's 20,038. This is particularly striking given that the Intel chip features 24 cores, suggesting a substantial leap in per-core efficiency for Nvidia's ARM implementation.
A Shift Toward Platform Integration
This move represents a strategic pivot for Nvidia, transitioning from a component supplier to a full platform provider. By integrating an ARM CPU with its own Blackwell-architecture graphics, Nvidia is attempting to replicate the vertical integration strategy used by Apple with its M-series chips. The leaked SoC integrates a Blackwell GPU featuring 6,144 CUDA cores. While these are engineering samples, current estimates place the GPU performance near the level of a GeForce RTX 4060.
Industry Implications
If the RTX Spark can maintain this level of performance while operating within the Windows 11 ARM ecosystem, it could disrupt the high-end mobile workstation and gaming segments. The primary advantage lies in the unified memory architecture and the synergy between the CPU and the Blackwell GPU, which could offer superior power efficiency and data throughput compared to the traditional split between a third-party CPU and a discrete Nvidia GPU.
What to Watch
While the raw benchmark numbers are promising, the real-world success of the RTX Spark will depend on software compatibility and driver stability for Windows on ARM. It remains to be seen how the final retail silicon will compare to the engineering samples in terms of thermal management and sustained clock speeds. Industry observers will be looking for official confirmation of the product lineup and the specific laptop partners Nvidia will choose for the initial launch.