NVIDIA has unveiled RTX Spark, a new platform built that brings personal AI to Windows PCs. Co‑developed with Microsoft and MediaTek, it combines gaming, AI, and portability in one chip, positioned to rival Apple, Intel, AMD, and ARM‑based laptops.
The chip is built on a 3nm Arm architecture with dynamic power scaling. It can run at single‑digit watts for light tasks and scale up to 80W for heavy rendering or gaming, keeping performance consistent whether plugged in or on battery.
At its core, RTX Spark integrates a 20‑core Grace CPU and Blackwell GPU, sharing up to 128GB LPDDR5X RAM through NVLink‑C2C at 300 GB/s. This unified setup removes bandwidth bottlenecks and ensures smooth multitasking.
For creators and gamers, it can handle 90GB+ 3D scenes, decode 12K raw video, and push AAA titles past 100 FPS at 1440p with DLSS 4.5
On the AI side, it delivers 1 petaflop of compute power, capable of running 120B‑parameter language models offline, with privacy safeguarded by NVIDIA OpenShell.
NVIDIA makes it clear this isn’t for casual users. RTX Spark is overkill for spreadsheets or streaming. It’s built for professionals who need heavy rendering, coding, or gaming without bulky adapters.
Also Read: NVIDIA unveils DLSS 5 with AI-powered graphics
Slim laptops (around 14mm, 3 lbs) with OLED G‑SYNC displays and compact desktops will launch this fall (Q3/Q4 2026) from ASUS, Dell, HP, Lenovo, Microsoft Surface, MSI, Acer, and GIGABYTE.
