RTX Spark Debut: HP’s $3,200 OmniBook Ultra 16 Beats Snapdragon, Lags M5

RTX Spark Arrives: What HP’s OmniBook Ultra 16 Packs

Nvidia’s RTX Spark platform is finally in retail laptops, and HP’s OmniBook Ultra 16 is the first out of the gate. It’s a thin-and-light 16-inch machine aimed at creative pros and AI developers, not gamers first.

The OmniBook Ultra 16 uses Nvidia’s new RTX Spark N1X system-on-chip, which combines an Arm-based Grace CPU with Blackwell RTX graphics and up to 128GB of unified memory. That’s a direct shot at Apple’s MacBook Pro and Qualcomm’s Snapdragon X Elite laptops.

HP offers only the higher-end N1X variant: a 20-core Grace CPU and a 6,144-core GPU. The lower-end 18-core version with 5,120 CUDA cores is reserved for HP’s OmniBook X 14, which starts at $3,000.

The design is all-aluminum and premium, with a 3K OLED touch display and a large haptic touchpad. HP also redesigned the hinge so the display rests on top of the back edge, leaving the rear vents unobstructed for better airflow. Two fans and three heat pipes handle cooling.

But power comes at a price. The base model with 32GB RAM and 1TB SSD costs $3,200. Doubling both pushes it to $5,000, and maxing out at 128GB RAM reaches $6,200. That’s MacBook Pro territory—or beyond.

Benchmarks: Strong Multi-Core, But Apple’s M5 Still Leads

In early Geekbench 6 and Cinebench 2024 runs, the RTX Spark laptops finished behind Apple’s latest 16-inch MacBook Pro with M5 Pro and M5 Max chips on both single- and multi-core tests.

Against Qualcomm’s Snapdragon X2 Elite, however, the picture is mixed. On Geekbench 6 multi-core, the RTX Spark systems performed similarly to the Asus Zenbook A16 and HP OmniBook Ultra 14, both of which use Snapdragon X2 Elite. But single-core performance was lower on the RTX Spark machines.

On Cinebench 2024, the RTX Spark laptops edged ahead of the Snapdragon X2 Elite systems. CNET tested three RTX Spark laptops—the HP OmniBook Ultra 16, Microsoft Surface Laptop Ultra, and Lenovo Yoga 9N 2-in-1 16 Gen 11—and they finished in a pack on most benchmarks.

Local AI workloads are where RTX Spark is expected to shine, thanks to the unified memory architecture and up to 300GB/s memory bandwidth. But the baseline 32GB configuration limits how much RAM you can dedicate to the GPU—only up to 16GB via the Nvidia app slider. For serious local AI agents, you’ll want the 64GB or 128GB models, which allow up to 48GB and 96GB GPU allocation, respectively.

The Arm-based architecture also means potential Windows-on-Arm compatibility issues. Many apps have native ARM versions, but others won’t run or will need Microsoft’s Prism emulator, which can hurt performance.

Design and Thermals: Thin, Loud, and Hot

The OmniBook Ultra 16 is thin and elegant, but it gets hot and loud under sustained load—just like a MacBook Pro.

HP’s new thermal design keeps the rear edge clear for airflow, and the all-aluminum chassis feels rigid and premium. The 3K OLED touch display is stunning, and the haptic touchpad is lively. The keyboard has a plush yet solid feel, though the lack of a number pad might annoy data scientists.

But even the most valuable company in the world can’t change thermodynamics. When pushed, the OmniBook Ultra 16 gets hot and loud, just as Apple’s similarly thin MacBook Pro does with its M-series silicon. It’s the price of packing desktop-class power into a portable frame.

Availability and Pricing

The HP OmniBook Ultra 16 is available now in the US, starting at $3,200 for the base configuration with 32GB RAM and 1TB SSD. The 64GB/2TB model is $5,000, and the 128GB model is $6,200. HP has not announced UK or EU pricing or availability yet, but given the US pricing, expect a direct conversion plus local taxes when it does arrive.

Other RTX Spark laptops from Asus, Dell, Lenovo, Microsoft, and MSI are also launching, with the Microsoft Surface Laptop Ultra and Lenovo Yoga 9N 2-in-1 16 Gen 11 already in testing. Availability will vary by region.

Nvidia’s RTX Spark is a serious entry into the high-performance laptop space, but it’s not a slam dunk. HP’s OmniBook Ultra 16 delivers strong multi-core and local AI performance in a premium package, yet Apple’s M5 Pro and M5 Max still hold the benchmark crown. For creative pros and AI developers who need CUDA and unified memory, the RTX Spark platform is worth a look—provided you can stomach the price and the Arm compatibility caveats.