Nvidia N1X Chip Specs Leaked: 20 Cores, Blackwell GPU, 128GB Memory

Nvidia is about to enter the Windows-on-Arm laptop market in a way that is very different from anything the company has done before. The tech industry has been closely watching this project, as shown by the specs that were leaked by VideoCardz a day before the expected announcement.

The N1 line has at least four different models, ranging from the high-performance N1X to the thinner, more mainstream N1 models that use less power. The whole line-up combines Arm CPU architecture with Blackwell 2.0 graphics in a single package.

This is Nvidia's clearest message yet that it wants to compete in the PC processor market as a full-stack silicon provider, not just as a GPU supplier for someone else's platform.

In a few days, the ban should be lifted. If the leaked specs are accurate, what Nvidia is releasing is a real architectural change from how the company has traditionally handled the laptop market.

The N1X: Flagship Specs That Rival Desktop-Class AI Hardware

The most powerful N1X configuration doesn't just use Nvidia's current chips; it's said to have the same core design as the GB10 processor that's in Nvidia's desktop AI supercomputer, the DGX Spark. That family tree is not a coincidence. It tells you what Nvidia is aiming for and who their main customer is.

The top-of-the-line N1X is said to have a 20-core CPU with ten Cortex-X925 performance cores and ten Cortex-A725 efficiency cores. This is a common hybrid configuration found in current high-end Arm designs.

A Blackwell 2.0 GPU with 48 Streaming Multiprocessors, which is the same as 6,144 CUDA cores, handles graphics. Up to 128 GB of LPDDR5X memory can be used through a 16-channel interface, and up to three M.2 SSDs can be used to add more storage space.

Another version of the N1X is also being planned. It will have 18 CPU cores arranged in a 9+9 pattern and a 40-SM GPU with 5,120 CUDA cores. Both N1X versions work within a 45W to 80W power range, which puts them in the same thermal class as today's top-of-the-line laptop processors.

The important difference is that this range includes both the CPU and GPU, not just the CPU. That's not the same kind of speed that TDP numbers for laptop chips usually mean.

At 128 GB of memory, you should take a moment to think. For most people, mainstream laptop computers from Apple, Qualcomm, and Intel are limited to 32–64 GB.

A laptop chip that can hold 128 GB is aimed at AI tasks, professional content creation, and developer needs where large model weights or complicated files need to be stored in a single memory. This chip isn't just for games; it also runs office programs. It's an AI desktop chip that does everything.

The Standard N1: Targeting Mainstream Laptops Without Abandoning Performance

The N1X is made for workstations and power users, but the normal N1 line is made for more general use. These devices are thinner, use less power, and are more popular with businesses.

The papers that were leaked show two different versions of the N1. The more powerful model has a 20-SM GPU with 2,560 CUDA cores and eight Cortex-X925 speed cores and four Cortex-A725 efficiency cores.

The second version comes with a 16-SM graphics processor and 2,048 CUDA cores, as well as a 10-core configuration with seven speed cores and three efficiency cores.

The power range for both N1 models is 18W to 45W, which is the range for most thin and light computers. Up to 64 GB of LPDDR5X memory can be used through an 8-channel link, and up to two M.2 SSDs can be used for storage.

The difference in power and memory between the N1 and N1X families is clear and planned. It looks like Nvidia is making a product stack where each level is designed for a specific type of customer, rather than a confusing middle ground.

While still in the lower end of the N1 range, having 2,048 CUDA cores in a laptop package that uses less than 45W is a big improvement over what Intel or AMD can offer at the same power level right now.

What a mainstream laptop GPU looks like could be changed by the N1's graphics performance for people who need GPU acceleration for AI inference, video editing, machine learning processes, or even high-fidelity gaming.

The Arm Architecture Bet and What It Means for Windows Compatibility

Nvidia chose to use Arm CPU architecture instead of x86, which puts it in the same architectural group as Apple Silicon and Qualcomm's Snapdragon X series. These are the platforms that have shown Arm can provide competitive speed and battery life on Windows and macOS.

The Windows-on-Arm community has grown up a lot since its early, rough days. Microsoft put a lot of money into making native Arm program support possible.

Now, most developer tools, Adobe, and Microsoft's own productivity suite can run natively on Arm without slowing down the system. This better fit is what makes 2025 and 2026 a more likely time for a new Windows-on-Arm app than earlier years.

It's not Nvidia's CPU design that gives it an edge in this market; Qualcomm's Cortex cores and Apple's custom silicon are both good in that area. The graphics processing unit (GPU). Blackwell 2.0 integrated graphics running at 48 SMs is a big step up from any other integrated graphics option on the laptop market right now.

For people who need GPU compute—and that group is growing quickly as AI-assisted apps become more common—NVIDIA's N1X offers something that no other Windows laptop chip does at the same level of speed.

The CUDA environment is the moat that goes deeper. The standard development platform for AI and GPU-accelerated computing around the world is Nvidia's software stack, which includes CUDA, TensorRT, and cuDNN.

If a laptop has native CUDA support, writers can run the same code that works on hardware in a data center on that laptop, without having to change or tweak the code. That's a strong reason for AI developers and researchers to think about getting a N1X machine, no matter what Apple Silicon or Qualcomm has to offer.

Development Timeline and What the 2024 Document Date Suggests

One thing that's easy to miss in the leaked documents is an important one: one of the internal slides supposedly has a date of 2024 on it, which means that the N1 project has been in the works for at least two years.

That timeline fits with what Nvidia has said in public about its goals for a wider range of platforms. The company has been working for a while to create a more complete computing stack that goes beyond separate GPUs, and the N1 family seems to be the laptop version of that strategy.

The fact that the product has been in development for two years also suggests that the specification leak is of a fairly developed product, not an early prototype that might not show how it will be shipped. It is important to pay attention to the caveat in the VideoCardz report: not all of the models mentioned are guaranteed to make it to market, and some configurations may not show up in shipping products.

But the four-model family, which includes both high-performance workstations and mainstream thin-and-lights, shows that this isn't just a limited first release, but a real product plan.

Since the embargo is expected to be lifted within days of the leak, the official information about specs, price, and availability should be coming out soon. What was leaked proves the system. When the start is announced, it will be clear which OEM partners are making the N1X and N1 laptops, how much they cost, and when they will be available.

Who This Chip Is Actually For

The N1X is clearly aimed at AI developers, researchers, and technical workers who need the most GPU power in a portable package. The maximum memory size of 128 GB and the number of CUDA cores that it has are too many for games or video editing. They're meant to run local AI models, fine-tune workflows, and do GPU-accelerated computations that used to need a separate GPU or desktop workstation. The fact that the chip design comes from DGX Spark makes this clear.

The standard N1 is made for a wider range of people, including pros who need a powerful GPU without the heat issues of a workstation, developers who need Arm compatibility and decent GPU acceleration, and consumers who want top-notch performance in a thin laptop. The 18W–45W range covers most of the general high-end laptop market. The MacBook Air and Pro are currently the most popular Arm laptops among Arm users.

The competitive tone makes people in both sides feel uncomfortable. Apple has been the leader in high-performance small laptops with M-series chips. Qualcomm's Snapdragon X has helped it become a leader in Windows on Arm.

Intel and AMD are still standing up for x86. Nvidia's move to sell both CPUs and GPUs along with the most popular GPU computing ecosystem gives Apple, Qualcomm, and x86 chipmakers a new competitor with structural benefits that none of them fully share.

The Bottom Line

With the N1X and N1 chips, Nvidia is making its most serious move yet to own a real piece of the laptop silicon market, not just as a GPU supplier for someone else's platform, but as the platform itself. There are four different models, ranging from 18W laptops for regular use to 80W workstations. All of them have Blackwell 2.0 graphics, up to 128 GB of LPDDR5X, and native CUDA support in a compact package.

The specs that got out suggest that Nvidia has really made something different. It's not a product that competes with Apple's Silicon or Qualcomm's Snapdragon X, but a chip that offers something that neither of them do: world-class GPU computing in a laptop package that runs the software stack that the AI industry already uses.

When the embargo ends, official price, OEM partnerships, and release dates will be made public. At least the design can be seen now, and it's more ambitious than any new laptop chip in years.