At Computex Taipei 2026, NVIDIA‘s CEO took the stage and officially unveiled the RTX Spark, the company’s first-ever processor built specifically for Windows PCs. It’s an Arm-based chip that packs a 20-core Grace CPU (built alongside MediaTek), a Blackwell RTX GPU with 6,144 CUDA cores, an NPU, and up to 128GB of unified LPDDR5X memory, all on a single TSMC 3nm package. 

As a result, NVIDIA says the RTX Spark is capable of 1 petaFLOP of AI performance. That’s a lot of zeroes.

The chip, formerly rumored under the codename “N1X,” is clearly NVIDIA’s answer to Apple silicon, and a direct challenge to Qualcomm’s Snapdragon X series that’s been dominating the Windows on Arm conversation for the past year. 

And NVIDIA isn’t just bringing its hardware here. The full software stack, including CUDA, TensorRT, DLSS 4.5, Reflex, G-SYNC, and RTX ray tracing, is all coming along for the ride.

What is NVIDIA RTX Spark capable of?

So what can it actually do? NVIDIA says the RTX Spark can run 120 billion parameter AI models locally, handle 12K 4:2:2 video editing, render 3D scenes larger than 90GB, and push AAA games at 1440p above 100 frames per second. The latter is powered by DLSS 4.5 and Frame Generation, which NVIDIA has been shipping on its desktop cards for a while now.

Adobe is on board too, and not in a half-hearted way. The company is doing a complete architectural overhaul of both Photoshop and Premiere to support RTX Spark, promising up to 2x improvements in AI and graphics performance. Blackmagic Design, Blender, CapCut, ComfyUI, and OTOY have also signed on.

The laptops powered by the RTX Spark will be slim and trim. NVIDIA says 14mm of thickness, 3 pounds of weight, and availability in 14-inch and 16-inch sizes, with OLED displays and precision-machined aluminum chassis. 

The first PCs with the new chip will launch this fall from brands like ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI. 

One thing that’s missing is the actual benchmark numbers. It’s still a few months until fall, but we can expect leaks revealing some performance figures before that. 

For more daily updates, please visit our News Section.

Stay ahead in tech! Join our Telegram community and sign up for our daily newsletter of top stories! 

The post NVIDIA RTX Spark is here, and it wants to be the Apple Silicon moment for Windows PCs appeared first on Gizmochina.

Nvidia is about to announce its first Arm chips for Windows laptops at Computex tomorrow. However, someone may have spoiled the surprise for Nvidia. 

Just a day before the expected announcement, detailed specifications of Nvidia’s upcoming N1-series Arm processors for PCs and laptops have surfaced online, courtesy of VideoCardz, which cited internal Nvidia documents.

The leak reveals at least four chip variants in the N1 family, covering everything from high-performance workhorses to power-efficient options for thin, mainstream laptops.

Some context first

For quite a while now, the Windows-on-Arm story has mostly been Qualcomm’s story. The company’s Snapdragon X Elite chips, built on a 4nm process with Qualcomm’s own Oryon CPU cores, brought real competition to Intel and AMD for the first time in years on thin-and-light laptops. 

Qualcomm made big claims about beating Intel Core i7 chips on multi-threaded benchmarks, and for the most part, the real-world results backed that up.

Now Nvidia wants in. And Nvidia isn’t just bringing a CPU to the fight; it’s bringing a Blackwell GPU along for the ride.

Nvidia N1X

The top-tier N1X reportedly shares its core design with the GB10 processor inside Nvidia’s DGX Spark desktop AI supercomputer. That’s not a casual comparison. The full-fat N1X is said to pack a 20-core CPU — ten Cortex-X925 performance cores and ten Cortex-A725 efficiency cores — paired with a Blackwell 2.0 GPU featuring 48 Streaming Multiprocessors, which works out to 6,144 CUDA cores.

A slightly trimmed N1X variant is also in the works, dropping to 18 CPU cores (nine performance, nine efficiency) and a 40-SM GPU with 5,120 CUDA cores. Both N1X chips are designed to run at 45W to 80W. Crucially, though, that power figure covers the entire chip package: CPU and GPU together.

Nvidia N1

The standard N1 lineup is for thinner, more affordable devices. Two variants have reportedly been planned. The higher-end one pairs eight Cortex-X925 performance cores and four Cortex-A725 efficiency cores with a 20-SM GPU delivering 2,560 CUDA cores. 

The second, more entry-level option steps down to a 10-core CPU (seven performance, three efficiency) and a 16-SM GPU with 2,048 CUDA cores. The whole N1 family runs within an 18W to 45W power envelope.

Memory and Storage

The two families differ significantly under the hood. The N1X supports up to 128GB of LPDDR5X memory across a 16-channel interface, while the standard N1 caps at 64GB with an 8-channel setup. 

Likewise, N1X supports up to three M.2 SSDs, while the N1 tops out at two.

How Long Has This Been Cooking?

According to VideoCardz, at least one of the leaked slides is dated 2024, meaning Nvidia may have been working on this for two years or more. Not every chip listed in these documents is guaranteed to ship; roadmaps change, and some variants may quietly disappear. But the scale of what’s been leaked suggests Nvidia had serious, long-term ambitions here.

For more daily updates, please visit our News Section.

Stay ahead in tech! Join our Telegram community and sign up for our daily newsletter of top stories! 

(Source)

The post Nvidia’s first ARM chip for PCs, N1 and N1x, leaks and it looks like a big deal appeared first on Gizmochina.