Nvidia RTX Spark laptops are here — and they’re not just another AI PC refresh. They’re a new hardware platform, built from the ground up on Arm, with the Grace Blackwell architecture, unified memory, and NVLink-C2C interconnect baked in. The first models ship with the N1X system-on-chip, fabricated by TSMC on a 3-nanometer process, and scale between 45W and 80W TDP. That lets them fit into laptops as thin as 0.55 inch and light as 3 pounds, while still delivering petaflop-class performance for local AI workloads. Nvidia says battery-powered operation retains 99% of plugged-in performance. That’s not incremental. It’s architectural. And it’s shipping now. Not next year. Not ‘early access’. Now. But only to developers and enterprise buyers. At least for launch. No consumer pricing or retail availability has been announced. Not in the US. Not in India. Not anywhere. Just confirmed specs, confirmed architecture, and confirmed deployment in early production units. This is not a Copilot+ laptop. It’s not an x86 gaming rig with a bigger GPU. It’s something else entirely: a mobile DGX Spark. A workstation that fits in a backpack. A device built for agentic AI, not just inference. It runs Windows on Arm, but unlike Qualcomm’s Snapdragon X Elite systems, it pairs Arm CPU cores (Cortex-X925 and Cortex-A725) with a full Blackwell GPU and up to 128GB of coherent unified memory. No PCIe bottleneck. No memory duplication. Just one memory space, shared across CPU and GPU, with 600GB/s of NVLink-C2C bandwidth. That changes what ‘local AI’ means. You’re no longer limited to tiny quantized models or background tasks. You can run multi-billion-parameter models with context retention, multitask AI workloads alongside other apps, and do it without throttling or swapping. Microsoft and Nvidia aren’t calling this a ‘laptop upgrade’. They’re calling it a new category. And the first machines are already rolling off production lines. Aimed squarely at AI engineers, ML researchers, and cloud-averse dev teams. No release date for broader markets. No list price. No Indian or US MSRP. Just silicon, specs, and a shift in what a Windows machine can be.
Nvidia RTX Spark laptops are here: A new Windows platform for agentic AI
The RTX Spark isn’t a new GPU or a new laptop line. It’s a new Windows platform — built on Arm, powered by Grace Blackwell, and designed for agentic AI workloads on the go.
By Clock Speed
The Computing Desk · (2 hours ago)

Reported from
How this story was made
Written by the Hitechreports desk from the reporting credited above, with facts attributed to their original publishers. We do not test devices ourselves; anything about performance, battery life or cameras comes from the outlets that did. Prices are as reported at the time of writing. Editorial policy · Report an error
Filed by The Computing Desk
Laptops, desktops, processors, graphics cards and the operating systems that run on them.
More from this desk →Read next
More Laptops & PCs →
Nvidia RTX Spark laptop: same performance on battery as on AC
Nvidia RTX Spark laptops are shipping now — and they’re the first Windows PCs built on Arm that promise identical performance whether plugged in or running on battery.

Microsoft Tunes Windows for Nvidia RTX Spark Laptops With Custom Emulation and Anti-Cheat
Microsoft has reworked Windows for Nvidia's Arm-based RTX Spark laptops, introducing a two-click agent onboarding process, tuned x64 emulation, anti-cheat gaming support, and unified memory management sliders.

Nvidia RTX Spark Laptop Review Shows On-Device Agents Running Complex Workflows
Early testing of Nvidia's RTX Spark superchip shows how on-device AI agents handle OS settings and creative tasks in DaVinci Resolve, backed by up to 128GB of unified memory.
Be the first to comment
Join the argument. No password, just your email or a passkey.