What is Nvidia's N1X chip? 11 questions and answers on the Windows AI PC, the Microsoft partnership, CUDA and Australian prices
Introduction: what does one chip change for the Windows PC?
This explainer brings together eleven questions and clear answers about Nvidia's N1X chip: what it is, when it arrived, where it fits in the long Nvidia and Microsoft partnership, the lessons of NV1, DirectX and 3dfx, how CUDA, Tensor Cores and NVLink shape modern AI, and what the first N1X laptops cost in Australia.
1. What is Nvidia's N1X?
Nvidia's N1X is a high-speed System-on-a-Chip (SoC) built for Windows 11 laptops (RTX Spark), creator laptops and small desktop computers. It combines a 20-core ARM Grace CPU, developed together with MediaTek, and a Blackwell RTX GPU with 6,144 cores on a single piece of silicon. With 128 GB of unified memory, it can run large AI models (local LLMs) directly on the computer, without an internet connection.
2. When was the N1X chip released?
The N1X chip was introduced on May 31, 2026, and officially went on sale in laptops on October 16, 2026. Small desktop computers followed in November 2026.
3. Does it work only on Windows computers, or on Mac as well?
It is built specifically for Windows 11 PCs only; it does not run on Mac computers. Apple uses only its own M-series chips with macOS. For Linux on N1X machines, Nvidia recommends using the Windows Subsystem for Linux (WSL).
4. What is the history of the Nvidia and Microsoft Windows partnership, by year, month and date?
September 30, 1995: Microsoft released DirectX 1.0 software for Windows 95, and Nvidia designed its RIVA 128 and GeForce chips accordingly. March 10, 2000: Microsoft and Nvidia signed a contract to build the graphics chip for the first Xbox game console. November 30, 2006: the two companies worked together for Windows Vista and DirectX 10 technology. October 26, 2012 and October 22, 2013: the Microsoft Surface RT and Surface 2 tablets were released with Nvidia's Tegra 3 and Tegra 4 chips. October 26, 2015: the first Surface Book laptop with a dedicated Nvidia GeForce GPU was launched. March 19, 2018: the two companies jointly introduced real-time ray tracing (DirectX Raytracing, DXR) on Windows 10. February 21, 2023: a ten-year agreement was signed to deliver Xbox PC games on Nvidia's GeForce NOW cloud platform. May 31, 2026 and October 16, 2026: the RTX Spark N1X platform for Windows 11 AI computers was introduced and went on sale.
5. When did Microsoft reject or sideline Nvidia?
Xbox 360 and Xbox One (2003-2013): after the first Xbox, a dispute over chip pricing pushed Microsoft away from Nvidia and towards ATI/AMD. Surface and the Qualcomm agreement (2014-2023): after the failure of Windows RT, Microsoft dropped the Tegra chips and held an exclusive agreement with Qualcomm for Windows on Arm laptops from 2016 to 2023. Data centre and Maia AI chips (2023-2026): to reduce its dependence on Nvidia GPUs alone, Microsoft built its own Azure Maia AI chips. Surface gaming laptop cancelled (June 2026): Microsoft abandoned a plan to build a standalone Surface gaming laptop, to avoid competing with its other Windows partners.
6. What happened in the 1990s with the NV1 failure, Microsoft's DirectX problem and the 30-day cash reserve?
NV1 and QTM technology (May 1995): Nvidia's first chip, the NV1, used curved surfaces (Quadratic Texture Mapping) instead of flat triangles. It was technically superior, but very hard for game developers to work with. Microsoft's DirectX problem (September 1995): the DirectX 1.0 (Direct3D) software Microsoft released made triangle-based graphics the common standard. New Windows games would not run on the NV1 chip, and Nvidia's technology became outdated overnight. NV2 cancelled and Sega's help (1996): Jensen Huang cancelled halfway the NV2 chip being built for Sega. With only enough money left for 30 days, he travelled to Japan, told Sega's CEO the truth, and received the remaining 5 million dollars of the contract as a donation, saving the company. The rise of RIVA 128 (1997): with that money Nvidia built the triangle-based RIVA 128 chip suited to DirectX, sold one million units in four months, and rose again.
7. How did Nvidia bring down 3dfx, and what are Jensen Huang's leadership qualities?
The fall of 3dfx: while 3dfx relied on separate 2D/3D cards and its own Glide software, Nvidia released the RIVA 128 (all-in-one) and the world's first GPU, the GeForce 256 (1999). By releasing a new chip every six months, Nvidia drove 3dfx into bankruptcy. Leadership qualities: the alert mindset that "we are only 30 days from going out of business", the honesty to openly admit mistakes (intellectual honesty), testing on a software emulator before making hardware, and a flat organization without many layers of titles are Huang's key leadership traits.
8. How did CUDA software turn gaming GPUs into the foundation of the AI revolution?
CUDA, introduced by Nvidia in 2006, made it possible to use the GPU's thousands of cores for scientific and mathematical calculations through C/C++ languages, without knowing graphics languages. In 2012, an AI model called AlexNet used CUDA on two Nvidia gaming cards to win the ImageNet competition. That historic victory became the starting point of today's modern AI revolution.
9. How do Tensor Cores and NVLink work in modern AI models (LLMs)?
Tensor Cores: instead of multiplying only two numbers like ordinary cores, they multiply an entire table of numbers (a matrix) at the same time and finish AI calculations at lightning speed with lower bit precision (FP8/FP4). NVLink: a high-speed data bridge that connects many GPUs directly, going beyond the speed limit of the motherboard (PCIe). It makes hundreds of GPUs work like one giant super chip.
10. How are Tensor Cores and NVLink used in the N1X chip in laptops?
The 5th-generation Tensor Cores in the N1X chip deliver 1 petaflop of AI speed and help run large AI models (local LLMs) without the internet. NVLink-C2C technology connects the CPU and GPU inside the same chip and makes the 128 GB of memory common to both, so the delay of copying data disappears and battery life improves.
11. Which major laptops come with the N1X chip, and what do they cost in Australian dollars?
Microsoft Surface Laptop Ultra: 15-inch screen, 18-core to 20-core, priced from AU$4,399 to AU$9,649. Lenovo Yoga Pro 9n / Yoga 9n 2-in-1: 16-inch 2.8K OLED screen, priced from AU$5,799 to AU$7,999. MSI Prestige N16 Flip AI+: 16-inch OLED touch screen, priced from AU$5,999 to AU$10,999 (128GB RAM). ASUS ProArt P16 (N1X): 16-inch 3K OLED screen, 48GB RAM, priced at AU$6,999. Dell XPS 16 Creator Edition: 16-inch 3.2K OLED screen, 20-core N1X, priced from AU$8,706.41.
Conclusion
The N1X carries three decades of Nvidia's history with Windows into a single chip: the lessons of NV1 and DirectX, the endurance of the Microsoft partnership, and the CUDA, Tensor Core and NVLink technologies that power modern AI. For buyers, the first N1X laptops make local AI a practical feature of an ordinary Windows PC, at Australian prices starting from about AU$4,399.