HP says the OmniBook Ultra 16 is the world’s thinnest laptop with NVIDIA RTX Spark. The 16-inch notebook packs up to 128 GB of shared memory, AI performance of up to one petaflop in FP4, and a 3K OLED display into a chassis that’s just 15.73 millimeters thin. An exact launch date and final pricing have not yet been confirmed – HP is planning availability for this autumn. At the same time, HP announced the OmniDesk, a compact desktop PC featuring the same chip.
The target audience is clearly defined: people who want to run local language models, experiment with agent-based workflows, or use computationally intensive creative software without relying on cloud services. “HP is focused on developing devices that integrate intelligent features directly into everyday workflows,” said Samuel Chang, Senior Vice President at HP, at the unveiling of the two devices.
Key Takeaways
- According to the manufacturer, the HP OmniBook Ultra 16 is the thinnest laptop with NVIDIA RTX Spark (15.73 mm)
- Up to 128 GB of LPDDR5X RAM and 1 petaflop of AI performance at FP4
- 16-inch, 3K OLED display with VESA DisplayHDR True Black 1000, up to 120 Hz
- 99-Wh battery for up to 17 hours of battery life; 50% charge in about 30 minutes
- Launch expected in autumn 2026, pricing not yet announced
- The HP OmniDesk, the desktop counterpart featuring RTX Spark, is expected to launch in April 2027
What exactly is NVIDIA RTX Spark?
NVIDIA RTX Spark is a new platform for Windows PCs that combines personal AI agents, local AI applications, and classic RTX graphics performance on a single chip. The platform is based on the so-called N1X superchip, which NVIDIA had already announced in detail ahead of gamescom 2026 and which will be featured in the first devices from partners such as Asus, Dell, Lenovo, MSI, Microsoft, Acer, and HP starting in October 2026.
The chip comes in two configurations: a variant with 20 CPU cores, 6,144 CUDA cores (Blackwell architecture), and 24 to 128 GB of unified memory for laptops and mini-PCs; and a smaller variant with 18 CPU cores, 5,120 CUDA cores, and up to 32 GB of memory, which will be used exclusively in notebooks. According to initial estimates, the GPU performance of the larger variant is roughly on par with that of a GeForce RTX 5070. Thanks to the generous, shared memory, however, it can also run larger AI models locally than a desktop graphics card with fixed GDDR7 memory would allow. Incidentally, RTX Spark should not be confused with the NVIDIA DGX Spark, a separate mini-computer for AI developers: RTX Spark is the version of the platform tailored for Windows laptops and desktops.
With RTX Spark, NVIDIA is competing directly against Intel, AMD, and Qualcomm in the AI PC market. The segment has been growing steadily since the first Copilot+ laptops with Snapdragon X Elite, and local NPU and GPU performance have long since become key selling points.
HP OmniBook Ultra 16: Detailed Specifications
The OmniBook Ultra 16 features a newly developed tower-hinge and trunk architecture designed to ensure better air circulation and heat dissipation within an exceptionally slim chassis. HP calls this a first for a consumer laptop. Larger heat pipes and two thin fans are designed to ensure that performance remains stable even during extended AI and creative workloads. Whether this can be achieved quietly in everyday use remains to be seen in initial tests.
| Component | HP OmniBook Ultra 16 |
|---|---|
| Processor/GPU | NVIDIA RTX Spark N1X, optional 18-core CPU / 5,120 CUDA cores or 20-core CPU / 6,144 CUDA cores |
| Memory | 32, 64, or 128 GB LPDDR5X-9400 (unified memory) |
| AI Performance | up to 1 petaflop in FP4 |
| Storage | Up to 2 TB PCIe Gen 5 NVMe SSD |
| Display | 16 inches, 3K OLED (2,880 × 1,800 pixels), up to 120 Hz, 500 nits SDR / 1,100 nits HDR, 100% DCI-P3, VESA DisplayHDR True Black 1000 |
| Battery | 99 Wh, up to 17 hours of battery life, 50% charge in about 30 minutes (140-W USB-C charger) |
| Weight/Thickness | approx. 2.2 kg, 15.73 mm |
| Ports | 2× USB-C (40 Gbps, DisplayPort 2.1, Power Delivery), 2× USB-A (10 Gbps), 3.5-mm jack, microSD, Wi-Fi 7, Bluetooth 6.0 |
| Other | Quad speakers with Smart Amplifier, 5-MP IR webcam, haptic touchpad |
To compare how RTX graphics performance typically translates to compact form factors, take a look at the ASUS ProArt GeForce RTX 5090: While that model packs significantly more raw power into a traditional desktop chassis, the OmniBook Ultra 16 offers the same RTX technology in a fraction of the space. Records for thin laptop chassis are nothing new anyway; even the 9-millimeter-thin Acer Swift 7 caused a stir back in the day. The difference is that the OmniBook Ultra 16 has to accommodate significantly more computing power and, consequently, greater cooling requirements.
HP OmniDesk: AI Computing Power for Your Desk
In addition to the laptop, HP also announced the OmniDesk, a compact desktop PC featuring RTX Spark. HP plans to release details on the chassis, ports, and exact specifications at a later date. The only thing that’s certain so far is that, thanks to its consistently available performance, the OmniDesk is designed to keep AI agents and background processes running for longer periods—even when no one is currently using the device. This represents continuous operation for AI tasks—a capability that has been largely absent from traditional Windows desktops until now.
Price and Availability
According to HP, the HP OmniBook Ultra 16 is expected to be available this autumn on HP.com and in stores. HP has not yet announced specific pricing – details are expected to be revealed closer to launch. The HP OmniDesk is expected to follow in April 2027 at the earliest; HP has not yet announced pricing for it.
Conclusion: Plenty of performance, but also a hefty price tag
The HP OmniBook Ultra 16 packs more local AI computing power into a chassis that was previously reserved for slim Ultrabooks. That’s the real news—not just the form factor itself. Whether the OmniBook Ultra 16 is worth it for the average user will depend on its final price – not yet announced – as well as how heavily they actually rely on local AI workflows instead of cloud services. Anyone who only chats with ChatGPT occasionally doesn’t need this. For anyone experimenting with large open-source models or looking to run computationally intensive creative software locally, the OmniBook Ultra 16 is already one of the most exciting RTX Spark laptops ahead of the platform’s official launch in October 2026.
Sources: HP Newsroom
