The NVIDIA N1x SoC represents a significant potential shift in PC architecture—integrating high‑end GPU performance within an efficient Arm‑based CPU package.
NVIDIA N1x Architecture:
NVIDIA and MediaTek collaborated to develop the N1X, an Arm architecture SoC designed for the Windows PC market. The family includes a lighter N1 variant for laptops and the N1X for desktops.
It has a large, 20-core CPU. 10× Cortex X925 performance cores and 10× Cortex A725 efficiency cores make up the minimal configuration.
NVIDIA N1x GPU:
The onboard GPU utilises NVIDIA’s Blackwell architecture with 48 Streaming Multiprocessors (SMs), for a total of 6,144 CUDA cores, the same as the number in a discrete RTX 5070 GPU. Tested at an average boost clock (~1.05 GHz), the OpenCL result on an early engineering sample was 46,361, which is higher than contemporary integrated GPUs—but far behind a fully powered RTX 5070 discrete GPU.
NVIDIA N1x Performance:
Leaked Geekbench 6 single-core and multi-core scores on Linux run around 3,096 and 18,837, respectively—only comparable with high-end laptop chips like Intel Core Ultra 9 285HX and AMD Ryzen AI MAX, although Apple’s M4 Max remains superior in raw performance.
Reddit commentary points out that on Linux, the single-core score is extremely high—Intel Core i9‑14900K levels—on significantly less power (even under 40–60 W).
NVIDIA N1x Launch Date:
Early expectations of a late‑2025 launch (e.g., Computex or CES) were revised.
Now, multiple sources report that N1X is pushed back to 2026, or maybe even to early Q1, because of silicon design issues, revision needs, and interdependencies like Microsoft’s upcoming Windows OS release (reportedly “Windows 12”). DigiTimes also surmises that NVIDIA and MediaTek might first target enterprise-class systems and postpone wider consumer availability until market conditions are better.
Image source: Geekbench
Impact of NVIDIA N1x:
If it were completely optimised in the realm of higher clock rates and better memory bandwidth, the N1X could deliver discrete-level GPU performance in one unified SoC—a serious rival to AMD’s Strix Halo APU and Apple’s M-series chips for gaming and AI purposes.
Even early samples already surpass most other integrated GPUs, with great potential—particularly in handheld AI gaming notebooks under a 100–120 W budget.
Specs
| Feature | NVIDIA N1x | NVIDIA GB10 |
|---|---|---|
| Market | Desktop/Laptop | AI Developer |
| CPU | 20 cores (10× Cortex‑X925 + 10× Cortex‑A725) | 20 cores (10× Cortex‑X925 + 10× Cortex‑A725) |
| GPU | Blackwell architecture, 48 SMs → 6,144 CUDA cores | Blackwell architecture, 48 SMs → 6,144 CUDA cores |
| Current Benchmarks | ~3,096 single-core, ~18,800 multi-core (Linux) | ~3,096 single-core, ~18,800 multi-core (Linux) |
| OpenCL GPU Score | ~46,361 (engineering sample) | ~46,361 (engineering sample) |
| Memory | Shared LPDDR5X (no dedicated GDDR) | Shared LPDDR5X (no dedicated GDDR) |
| Power (TDP) | ≤ 120 W total | 170W |
| Expected Launch | Q1 2026 (enterprise first, consumer later) | July 2025 |
The NVIDIA N1x SoC is a potential category-buster in PC design—blending top-of-class GPU performance in an optimised Arm-based CPU package. While these are early engineering leaks, they suggest a highly competitive processor that is capable of replacing the laptop segment, especially for AI and gaming workloads. That said, the actual launch comes in 2026, and much depends on final silicon qualification, OS readiness, and OEM support.
Also Read: Asus Ascent GX10 launch date announced! It comes with the Nvidia GB10 Blackwell superchip.
