Microsoft and Nvidia Reunite for Arm-Based Surface Laptop Ultra
Microsoft and Nvidia just announced the Surface Laptop Ultra, a 15-inch Windows laptop powered by Nvidia's new Arm-based RTX Spark superchip. This marks Microsoft's second attempt at an Arm+Nvidia flagship after the original Surface's $900 million write-off.
The RTX Spark chip, originally sold in Nvidia's DGX Spark mini-PC for AI developers, is now optimized for Windows 11. It packs up to 20 CPU cores, 6,144 GPU cores, and 128GB of unified memory. Lower-end configurations will ship with as little as 16GB, according to Nvidia's briefings.
Performance Claims: RTX 5070-Class Graphics, 1 Petaflop AI
Microsoft Surface boss Andrew Hill stated: "This is the most powerful thing we've ever made." Nvidia claims the RTX Spark delivers roughly RTX 5070 laptop-level graphics performance and up to 1 petaflop of AI compute. The chip also promises "all-day battery life."
Display and Design
The Surface Laptop Ultra features a 15-inch mini-LED touchscreen at 262 pixels per inch with 2,000 nits peak HDR brightness — "the brightest display we've ever shipped," per Microsoft. It includes the largest haptic trackpad Microsoft has ever put on a Surface. The laptop weighs under 4.5 pounds and comes in dark grey and silver.
Ports and Connectivity
Ports include USB-C, USB-A, HDMI, a full-size SD card slot, and a headphone jack. Microsoft has not disclosed speeds or version numbers for these ports. The blog post shows what appears to be three USB-C ports, but this is unconfirmed.
Windows on Arm Ecosystem Push
Microsoft and Nvidia have been collaborating for years to optimize Windows for Arm devices. The Surface Laptop Ultra will launch this fall alongside other RTX Spark laptops and mini-PCs. Microsoft's blog post touts developer support: "No walls. No compromises." and "This is Surface craft at its most considered."
What Developers Should Know
- The RTX Spark chip offers 128GB unified memory (configurable down to 16GB), ideal for large AI model inference and training.
- 1 petaflop AI compute means you can run models like LLaMA 3 locally at impressive speeds.
- Windows on Arm is finally getting serious hardware — but check your toolchain for Arm compatibility.
A Note on the Source
This article is based on The Verge's report. Final specs, pricing, and exact port configurations remain unannounced. Follow Nvidia and Microsoft for updates.
Next Steps
If you're an AI developer, start testing your workloads on Nvidia's DGX Spark (same chip) to prepare for the Surface Laptop Ultra. If you're a Windows on Arm skeptic, this might be the hardware that changes your mind.
