NVIDIA AI PC Chip Lags Apple 2 Years, N1X Matches M3 Level

According to analysis from AppleInsider, NVIDIA’s newly unveiled RTX Spark Superchip, introduced at Computex 2026 as the company’s flagship platform for AI-powered Windows PCs, may still trail Apple’s Apple Silicon processors by nearly two years in CPU performance based on currently available benchmark data.

During his Computex keynote, NVIDIA CEO Jensen Huang positioned RTX Spark as a breakthrough platform designed to redefine personal computing by enabling users to run large language models (LLMs), AI agents, and advanced content creation workflows directly on local devices. The chip combines a 20-core Grace CPU, a Blackwell GPU, fifth-generation Tensor Cores, and up to 128GB of unified memory, connected through NVLink-C2C technology that delivers up to 600GB/s of bandwidth.

NVIDIA stated that RTX Spark is capable of handling 90GB-scale 3D scene rendering, generating AI-powered 4K video content, supporting 12K video editing workflows, and running large language models with up to 120 billion parameters and one million-token context windows on-device.

AppleInsider noted that the overall architecture of RTX Spark closely resembles Apple Silicon’s integrated design philosophy, combining CPU, GPU, AI acceleration, and unified memory into a single package. The report suggests NVIDIA is attempting to replicate the strategy that helped Apple reshape the PC processor market with its Arm-based silicon.

However, questions remain regarding CPU performance.

The publication referenced leaked Geekbench benchmark results from an engineering sample of NVIDIA’s upcoming Arm-based N1X processor, which surfaced in 2025. The sample reportedly achieved a single-core score of approximately 3,096 and a multi-core score of 18,837.

For comparison, Apple’s M3 Max recorded around 3,128 points in single-core testing and approximately 20,969 points in multi-core performance. More recent Apple processors have widened the gap, with the M5 reportedly reaching 4,224 in single-core benchmarks and the latest M5 Max approaching 30,000 points in multi-core testing.

Based on these figures, AppleInsider argues that the N1X’s single-core performance is roughly comparable to Apple’s M3 generation, which debuted in 2023, while remaining behind the current M5 family. If final commercial products deliver performance similar to the leaked engineering sample, NVIDIA’s first Arm-based PC processor could lag Apple’s latest CPU technology by more than two years.

The report also cautioned that the benchmark data originated from an early engineering version evaluated roughly a year ago. Clock speed improvements, power efficiency optimizations, and architectural refinements could significantly alter the performance profile of the final product before commercial release.

Despite potential CPU limitations, AppleInsider believes NVIDIA’s strongest advantage lies in its extensive AI ecosystem. The company’s CUDA platform, Blackwell GPU architecture, and mature AI software stack provide a competitive foundation that could help RTX Spark gain traction in the rapidly expanding AI PC market.

While RTX Spark represents NVIDIA’s most ambitious effort yet to bring advanced AI computing capabilities to Windows PCs, current information suggests the platform may still face significant challenges in matching or surpassing the performance of Apple’s latest Apple Silicon processors. Nevertheless, NVIDIA’s leadership in AI hardware and software integration could prove to be a key differentiator as the AI PC industry continues to evolve.