Nvidia, the leading name in the graphics card market, is making a historic breakthrough that will shake things up in the computer world. Taking the stage at the Computex 2026 fair, Nvidia CEO Jensen Huang announced the company’s first official processor family, which has been subject to leaks for a long time and is known by the code names “N1” and “N1X”. Nvidia RTX SparkThis new “Superchip” series, called “Superchip”, shows that the company has entered the PC market as a direct competitor to Intel, AMD and Qualcomm.

These new Arm-based processors, developed in partnership with Microsoft and MediaTek, usher in a brand new era of artificial intelligence and gaming on both Windows laptops and compact desktop computers. RTX Spark, which will be available this fall, aims to transform computers from conventional devices into local artificial intelligence spies.

• Nvidia officially announced its first computer processor family, known as N1 and N1X in leaks, under the name “RTX Spark” at Computex 2026.

• Produced in collaboration with MediaTek, the Arm-based processor offers a 20-core CPU, integrated GPU with 6,144 Blackwell CUDA cores and 1 petaflop artificial intelligence processing power.

• The processor family, which will be used first in the Microsoft Surface Laptop Ultra and Dell XPS 16 models, will be released this fall.

The First All-in-One Chip Powered by Blackwell Architecture

Nvidia now brings the power it offers in data centers and flagship graphics cards to our computers in a single processor die. The model, which is the top version of the RTX Spark processors and is referred to as N1X in leaks, resembles a portable monster. This processor, which comes off the band with TSMC’s advanced 3nm production architecture, is formed by connecting two different chiplets to each other via a high-speed silicon bridge.

On the processor (CPU) side of the chip, there is a 20-core structure designed with the contributions of MediaTek. This structure consists of 10 high-performance-oriented Arm Cortex-X925 cores and 10 power-efficient Arm Cortex-A725 cores and can reach a clock speed of up to 4.1 GHz.

The real big surprise comes on the graphics (GPU) side. The integrated graphics unit is based on the company’s latest Blackwell architecture and features a full 6,144 CUDA cores. This value exactly matches the core count of a desktop GeForce RTX 5070 graphics card. The company supports this combined structure with a unified LPDDR5X memory architecture that can go up to 128 GB and offers a high bandwidth of 600 GB / s.

Energy Efficiency and Cool Running Performance

Such high graphics power can be expected to drain the battery in laptops within seconds, but Nvidia is deeply integrating with Microsoft to solve this situation. RTX Spark operates with a very flexible thermal design power (TDP) profile, ranging from 45W to 80W depending on the hardware configuration.

In order to be fully compatible with the Windows operating system, Microsoft is introducing “Workload Profile Scheduling” (WPS) technology specific to this chip. This technology enables the system to manage hardware resources with the highest efficiency from day one.

At the same time, the Microsoft Power and Thermal Framework (MPTF) is hardware integrated into the chip to ensure that the devices remain cool and quiet even under heavy load. In this way, computers offer a battery life that lasts all day and stay away from heating problems.

1 Petaflop AI Power and Spy Era

Nvidia RTX Spark is positioned not only as a gaming or performance chip, but also as the cornerstone of the artificial intelligence computers of the future. Thanks to the 5th generation Tensor cores on it, this processor can locally produce exactly 1 petaflop artificial intelligence (FP4) processing power. This power, which far exceeds the 40 TOPS NPU limit that Microsoft has set as a rule for Copilot+ computers, radically changes the computer experience.

According to Jensen Huang’s words at the launch, computers are no longer static devices that we just open and click on applications. RTX Spark can run advanced artificial intelligence models, including Nvidia’s own Nemotron models, locally on itself, without needing the internet at all. In this way, devices turn into autonomous artificial intelligence spies that organize things on behalf of the user. Great steps are also being taken on the software support side.

Adobe doubles the performance of local artificial intelligence tools by fully supporting this new chip in Photoshop and Premiere applications. In addition, popular tools and productivity applications for software developers are also optimized to provide local support for the Arm architecture.

What are the Technical Details of N1 and N1X Models?

Nvidia offers the RTX Spark family with different core configurations to address the different needs of the market. Variations in leakage processes called N1 and N1X provide users with different options focused on budget and performance.

Feature	High-End (Legacy N1X)	Midrange (Legacy N1X Sub)	Entry Level (Old N1)
Number of CPU Cores	20 Cores (10P + 10E)	18 Cores (9P+9E)	12 or 10 Cores
GPU CUDA Core	6,144 Blackwell Cores	5,120 Blackwell Cores	2,560 or 2,048 Cores
Maximum Memory	128GB LPDDR5X (256-bit)	128GB LPDDR5X (256-bit)	64GB LPDDR5X (128-bit)
Power Consumption (TDP)	45W – 80W	45W – 80W	Up to 45W
Connection Support	PCIe 5.0 and PCIe 4.0	PCIe 5.0 and PCIe 4.0	PCIe 5.0 and PCIe 4.0

While entry-level models target thin and light Ultrabook models with lower power consumption, high-end models target performance-oriented portable workstations and gaming computers.

Emulation Question Mark in the World of Windows on Arm

Although this breakthrough by Nvidia is exciting, x86 application compatibility, which is the biggest handicap of the Arm architecture, comes to mind here as well. Especially game lovers wonder, “How will old games work?” In response to the question, Nvidia and Microsoft state that the “Prism” emulator in Windows is fully optimized for RTX Spark processors.

Old applications and games with 32-bit and 64-bit x86 architecture can run without any problems thanks to this emulation layer.

However, it is a known fact that emulation processes always cause some compromise in raw performance. Although Nvidia promises to close this gap with the muscle power of the graphics processor, the actual performance of the chip in games will become clear with detailed tests to be carried out this autumn.

The new processor family first appears in Microsoft’s new flagship laptop, Surface Laptop Ultra and Dell XPS 16 models. Subsequently, big names in the sector such as Asus, Lenovo and HP are preparing to launch their new generation RTX Spark-based computers.