Science & Tech · Gadgets & EVs · 1 hr ago
Nvidia RTX Spark laptops target local AI with up to 128GB of shared memory
Nvidia’s N1X chip powers the new RTX Spark laptops designed for demanding tasks, including running AI models on the computer itself.
Reviewers tested HP’s OmniBook Ultra 16 and Asus’s ProArt P14, two different laptops using the platform.
Their shared memory can reach 128GB, giving the graphics and central processors access to a large pool of fast memory.
That matters because local AI models can need far more memory than ordinary laptop graphics hardware provides.
The HP model performed strongly in local AI tests, and its reviewer said few Windows notebooks would be faster for this work.
Prices start at about $2,600, putting these machines well above the price of most laptops.
They may suit AI researchers, enthusiasts and creative professionals, while the Asus review also examined gaming and Linux use.
Nvidia’s RTX Spark laptops are designed to run demanding AI models locally, with up to 128 GB of shared memory.
The first RTX Spark laptops began shipping and reaching stores in mid-October 2026.
The Register found HP’s OmniBook Ultra 16 delivered very strong local AI performance, but said prices start around $2,600 and can exceed $6,000.
Reviews of several models praised their performance for creators, while some raised questions about local AI’s readiness and battery life.
The laptops use Nvidia’s N1X superchip, which combines CPU and GPU capabilities.
- Who
- Nvidia and laptop makers including ASUS, Dell, HP, Lenovo, Microsoft and MSI. Reviewers tested the first RTX Spark laptops.
- What
- The first RTX Spark laptops are arriving, with up to 128 GB of unified memory intended to support local AI workloads.
- When
- The first models began shipping and reaching stores in mid-October 2026.
- Where
- The laptops are being sold by participating manufacturers and retailers. The articles do not specify a launch market.
- Why
- The shared memory can support larger local AI models than systems with less video memory. The laptops also target AI enthusiasts, researchers and creative professionals.
This story does not have two clearly opposing sides.
Apple began using unified memory in its M-series chips, according to Heise.
Nvidia promoted RTX Spark laptops at Computex, according to Heise.
The first RTX Spark laptop models began shipping and reaching store shelves.
The Register published its benchmark review of HP’s OmniBook Ultra 16.
Windows Central reported that reviews of many first-wave RTX Spark laptops had gone live.
- Memory
- Up to 128 GB of unified memory
- Starting price
- Around $2,600
- Top configurations
- More than $6,000
- HP review model
- OmniBook Ultra 16
- HP display
- 16-inch, 2880×1800 OLED touchscreen
- Heise review model
- ASUS ProArt P14
3 sources
What people are saying
you'll be hard pressed to find a faster Windows notebook for local AI.
It's a laptop for those who know they need it, and it's mainly for AI developers and high-end content creators.
It all feels like a work in progress.