| CPU Configuration | GPU Configuration | Power Budget (Watts) | Memory (according to leak) | |
|---|---|---|---|---|
| Nvidia N1X 675 | 20 cores, 10+10 | 48 SMs / 6,144 CUDA cores | 45–80 | up to 128 GB LPDDR5X |
| Nvidia N1X 650 | 18 cores, 9+9 | 40 SMs / 5,120 CUDA cores | 45–80 | up to 128 GB LPDDR5X |
| Nvidia N1 | 12 cores, 8+4 | 20 SMs / 2,560 CUDA cores | 18–45 | up to 64 GB LPDDR5X |
| Nvidia N1 | 10 cores, 7+3 | 16 SMs / 2,048 CUDA cores | 18–45 | up to 64 GB LPDDR5X |
Strategic Positioning
Nvidia is not marketing the RTX Spark as a mere hardware upgrade, but rather views the chip as the foundation for the era ofAI agents
. Together with Microsoft, the OpenShell software environment was developed to run generative AI models with up to 120 billion parameters entirely locally on the PC.
With 1 petaflop (FP4) of AI computing power, the system breaks free from cloud constraints and token limits. Data remains on the device, a major selling point for businesses and privacy-conscious creators. For creatives, the new chip is expected to deliver enormous benefits:Adobe is porting rendering engines (e.g., for Premiere Pro and Photoshop) specifically for this architecture, which, according to the manufacturer, is expected to double performance. Large 3D scenes or video editing benefit massively from the 128 GB of shared memory, which was previously unthinkable in particularly compact notebooks.
Gaming Performance and Efficiency
Nvidia promises AAA gaming at 1440p with over 100 FPS, even while running on battery power. The graphics performance is said to be on par with a mobile RTX 5070, but with significantly reduced power consumption: The entire SoC is designed for a maximum Thermal Design Power (TDP) of 80 watts. By comparison, a dedicated RTX 5070 laptop GPU can consume up to 100 watts on its own, and that doesn’t even include the heat dissipation from the processor. Features such as ray tracing, DLSS (including the announced DLSS 4.5), and Reflex are natively supported.
The Critical Success Factors
Whether the RTX Spark will revolutionize the PC market in the long term will depend on three essential factors:
- The Windows-on-ARM hurdle:Qualcomm has taken the lead with the Snapdragon X Elite, but the success of RTX Spark will depend largely on software emulation. Native apps run fantastically; there’s no doubt about that. But what about older x86 applications, specialized drivers, and, above all, anti-cheat systems in PC games? For Spark to be a true success, all of this would need to run as flawlessly and efficiently as possible. While Nvidia emphasizes its close collaboration with developers, only in practice will we see just how high the performance really is.
- The price point:Since the memory is soldered directly onto the chip, affordable upgrade paths for end users are out of the question. Considering that the Linux-based sibling, the DGX Spark (in the Asus Ascent), starts at around 3,500 euros, the first premium notebooks are likely to land in the absolute high-end price segment. In other words, these laptops certainly won’t be cheap.
- Market environment:Analysts are forecasting a slight decline in the overall PC market for the second half of the year. Nvidia must therefore convince consumers and businesses that local AI agents are worth the hefty premium and are not just an
expensive toy for enthusiasts
.
My opinion
The Nvidia RTX Spark is a technical masterpiece and the most aggressive challenge yet to Intel and AMD’s x86 dominance in the Windows segment. For creators, software developers, and AI enthusiasts seeking maximum workstation power in a slim form factor, the chip will set the new benchmark this fall.
In terms of hardware alone, Nvidia seems to be leaving no stone unturned. However, to achieve lasting success in the PC market, the company must above all achieve high acceptance among potential customers. This is likely only possible if end users face as few compatibility limitations as possible in their daily lives.
The second key factor will be price. For the upcoming Spark laptops, I would estimate a price tag of well over 3,000 euros for the top-of-the-line configurations featuring the fully equipped Spark chip and 128 GB of unified memory.
To achieve lasting success in the market, significantly more affordable models will also be necessary. So far, four versions appear to be planned. Ultimately, the performance of the entry-level versions in particular will determine how widely accepted they will be by buyers.
Already read?SSD enclosure with fan in review: Finally a cool SSD, or just marketing?
It should be clear that Nvidia is pushing hard into the market and leveraging its market power to make Spark a success. As a heavyweight in the gaming segment, the company definitely has the right tools at its disposal to finally breathe life into the ARM architecture for gaming on Windows on ARM.
Ultimately, Spark could finally pave the way for ARM architecture on Windows, and the odds are good. However, the exorbitant prices for DRAM are likely to make the endeavor significantly more difficult.