Home > News

Nvidia RTX Spark release date window, specs, price estimates, and confirmed models – everything we know

Everthing there is to know about Nvidia's new RTX Spark chip
Last Updated on
Nvidia RTX Spark release date window, specs, price estimates, and confirmed models – everything we know
PC Guide is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Read More

Unveiled during Computex 2026, Nvidia’s new RTX Spark Superchip is being advertised as hardware that “reinvents Windows PCs” for a wide variety of users, from AI developers to creators and gamers. The ARM-based chip, code-named N1X and made with some help from MediaTek, will be available later this year in a selection of prebuilt systems.

It’s easy to make comparisons between RTX Spark and Apple Silicon (M-series), which brought ARM architecture to Apple hardware, rather than relying on Intel-built chips. Windows is defined by its x86 architecture and dominated by Intel and AMD processors for it, and this is exactly what Nvidia aims to shake up by introducing ARM into the mix.

RTX Spark release window & DGX Spark desktops

In its official blog post, Nvidia says RTX Spark will be available in Fall 2026, including the range of RTX Spark-powered laptops and compact desktops shown off during Computex 2026 in Taipei. No specific date has been mentioned, and we can expect to see the exact availability range depending on the manufacturer.

As for larger desktop tower-style products, these are not currently planned for RTX Spark. Instead, they will be reserved for the DGX Spark chip, also for Windows desktops. During his keynote, Nvidia CEO Jensen Huang described one desktop as “an incredible system” with a whopping 768GB of memory. Nvidia says these workstation models will be available in Q4 2026.

RTX Spark specs

SpecificationNvidia RTX Spark
ArchitectureARM-based CPU with Nvidia RTX graphics
CPU20-core NVIDIA Grace CPU
GPUNVIDIA Blackwell RTX GPU
CUDA CoresUp to 6,144
Memory128GB Unified Memory – 600 GB/s NVLink C2C
AI performanceUp to 1 petaflop of AI compute
Nvidia technologiesCUDA, RTX, DLSS, TensorRT, OptiX, Reflex, G-Sync
Other45W to 80W power consumption, “All-Day Battery Life”

Nvidia says its RTX Spark chip is capable of running 120-billion-parameter AI models locally, handling 12K video editing, rendering 3D scenes larger than 90GB, and running AAA games above 100 frames at 1440p. The gaming claim is most certainly with help from DLSS 4.5 and frame gen. Two games shown on stage were Forza Horizon 6 and 007 First Light.

Speaking about the chip, Jensen Huang said during his Computex keynote that “100% of Nvidia software stack runs” on its N1X chip, referring to the RTX Spark chip’s internal codename. A more power-efficient chip, N1, will also be available in the RTX Spark line-up. Both will certainly be attractive pieces of hardware for AI workflows among Windows power users.

Nvidia has not officially announced the cut-down specs of the N1 (non-X) chip, though a leak suggests it will feature the same 12 and 10-core Grace CPU designs with up to 2,048 CUDA Cores. The lower-profile chip is expected to support up to 64GB of memory and run with 18W to 45W power consumption.

What RTX Spark models have been confirmed so far?

Nvidia namedropped the following partners who will be developing RTX Spark-powered machines this fall: ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI. A handful of products were shown off on stage during its Computex reveal, and Nvidia says there’s more to follow from Acer and Gigabyte at a later date. 30 laptops and 10 mini desktops have been confirmed.

On the Nvidia website, we can see some specific models that will be launching later this year. Below are all confirmed RTX Spark laptops, as no specific desktop PC products have been named, though they were present on the Computex stage.

  • ASUS ProArt P16
  • Dell XPS 16
  • HP OmniBook X 14
  • Lenovo Yoga Pro 9n
  • Microsoft Surface Laptop Ultra
  • MSI Prestige N16 Flip AI+
Jensen Huang holding RTX Spark laptops at Computex 2026
Jensen Huang holding RTX Spark laptops at Computex 2026, credit: Nvidia

RTX Spark model price estimates

Neither Nvidia nor its partners have officially announced any prices for its RTX Spark-powered machines. However, current estimates suggest they won’t be cheap. Financial firm Morgan Stanley reports that the top N1X models will cost not far $3,000 USD, while cut-down N1 chip designs are expected to start at more than $1,000 cheaper.

“Our checks with PC brands at this year’s Computex suggest AI PCs with N1X will need to price at US$2,899, while N1 models will be priced at US$1,799.”

Credit: Morgan Stanley, via Max Weinbach

Such high prices are to be expected from such a revolutionary piece of tech – remember, Windows desktops have relied on the x86 architecture for over 40 years, and the introduction of ARM is a major shake-up. Either way, specific model pricing should vary to some degree, and there should be a wide range of options available at launch this fall.

Is RTX Spark a standalone device?

The short answer is no. While RTX Spark is indeed a standalone system-on-chip (SoC), much like Nvidia’s DGX Spark, you cannot buy the RTX Spark separately as a standalone processor. It’s not the same as buying an Intel Core or AMD Ryzen CPU, for example. It will only be available in prebuilt OEM designs.

Nvidia worked with MediaTek to develop the chip and with SK Hynix for memory, but has primarily partnered with Microsoft to bring the project to reality – introducing ARM on prebuilt Windows PCs for the masses. Nvidia has also partnered with Adobe specifically to improve AI-accelerated tools in its popular Premiere and Photoshop software.


3426
Would you buy a bare bones Steam Machine?

Would you buy a variant of the Steam Machine with no RAM and/or no SSD and install them separately to help save on costs?

About the Author

At PC Guide, Jack is mostly responsible for reporting on hardware deals. He also specializes in monitors, TVs, and headsets and can be found putting his findings together in a review or best-of guide.