The OmniBook Ultra 16 pairs Nvidia’s N1X platform with up to 128GB of shared memory, targeting developers and researchers running large AI models without the cloud.
Nvidia’s RTX Spark push is extending high-end AI and graphics hardware beyond the data centre, with HP’s OmniBook Ultra 16 presented as an early example of the category. The machine is aimed at builders, researchers and engineers who need local compute capacity for large models and advanced workloads, rather than relying on remote cloud services. According to The Register, the key attraction is access to up to 128 GB of unified memory, a scale that is unusual in a laptop and designed to keep sizeable AI and machine learning models running on the device itself.
At the centre of the platform is Nvidia’s N1X system-on-chip, which combines a CPU chiplet based on MediaTek design work with an Nvidia Blackwell graphics unit through NVLink Fusion interconnect technology, The Register reported. The most capable configuration is said to include 20 CPU cores and a GPU with 6,144 CUDA cores, putting it in the same core-count territory as an RTX 5070-class graphics processor while adding access to shared system memory. That design is meant to reduce the usual bottleneck that limits conventional consumer laptops when handling memory-hungry professional workloads.
HP has wrapped the hardware in a 4.89-pound all-aluminium chassis with an external hinge arrangement intended to move heat away from the keyboard and towards the rear exhaust area. The laptop also includes a 16-inch OLED touchscreen with a 2,880 x 1,800 resolution and a quoted peak HDR brightness of 1,100 nits, alongside Cirrus Logic audio amplification and a low-profile membrane keyboard, according to The Register. The result is a machine that appears engineered as much for sustained thermal management as for portability.
The trade-off is cost. The Register said RTX Spark systems start at about $2,600 and rise beyond $6,000 for the highest-memory versions, placing them firmly in specialist territory. Connectivity is comparatively generous, with two 40 Gbps USB4 ports, two 10 Gbps USB-A ports, HDMI and a microSD slot. Nvidia and Microsoft are positioning the platform for startup teams and developers building local agentic software stacks who want CUDA compatibility without depending on alternative or less familiar GPU ecosystems.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
