NVIDIA's N1X Chip for Windows on Arm Leaked

It has been revealed through multiple leaks that NVIDIA is preparing a new mobile processor series, “N1” and “N1X,” designed for Windows on Arm laptops. These chips are optimized for laptops and are based on the architecture of the desktop AI supercomputer DGX Spark, which was announced last year. According to presentation documents and retail listings obtained by foreign media outlets like VideoCardz and WinFuture, NVIDIA is collaborating with Microsoft and is expected to officially announce these processors at Computex 2026 in Taiwan.

Bringing DGX Spark Technology to Mobile The

DGX Spark, a compact AI-dedicated device, integrates a 20-core Arm-based CPU with NVIDIA’s Blackwell-generation GPU, featuring up to 128GB of unified memory (bandwidth of 273GB/s). Initially priced at $3,000, its cost has now risen to $4,699. The aim of the new N1X series is to bring this high processing power to a wider audience. According to leaked materials, NVIDIA has prepared four chips. The flagship model, N1X (675), combines a 20-core CPU with a GPU based on the Blackwell 2.0 architecture, featuring 48 streaming multiprocessors (SMs) and 6,144 CUDA cores. With a TDP ranging from 45 watts to 80 watts, it is well-suited for high-end gaming laptops and mobile workstations. However, considering the price range of the DGX Spark, it is unlikely to be offered at an affordable price. Instead, it is expected to be positioned as a professional product targeted at developers and creators rather than gamers. The lineup also includes the 18-core N1X (40 SMs, 5,120 CUDA cores, TDP 18–45W), the 12-core N1 (20 SMs, 2,560 CUDA cores), and the 10-core N1 (16 SMs, 2,048 CUDA cores). The lower-end models are aimed at mid-range laptops and mini PCs, but NVIDIA is likely to focus its competitive efforts on the high-end segment.

Four Chip Configurations The specific

specifications of the chips are as follows. Each chip adopts a heterogeneous architecture combining high-performance cores and high-efficiency cores: - N1X (20 cores): 10 Cortex-X925 (high-performance) cores + 10 Cortex-A725 (efficient) cores. GPU: 48 SMs, 6,144 CUDA cores. Memory: 16–128GB LPDDR5x (16 channels). PCIe Gen5: 12 lanes; Gen4: 5 lanes. TDP: 45–80W. - N1X (18 cores): 9 Cortex-X925 cores + 9 Cortex-A725 cores. GPU: 40 SMs, 5,120 CUDA cores. Memory: 8–64GB LPDDR5x (8 channels). PCIe Gen5: 8 lanes; Gen4: 3 lanes. TDP: 18–45W. - N1 (12 cores): 8 Cortex-X925 cores + 4 Cortex-A725 cores. GPU: 20 SMs, 2,560 CUDA cores. Memory: 8–64GB LPDDR5x (8 channels). PCIe Gen5: 8 lanes; Gen4: 3 lanes. TDP: estimated 18–45W. - N1 (10 cores): 7 Cortex-X925 cores + 3 Cortex-A725 cores. GPU: 16 SMs, 2,048 CUDA cores. Memory: 8–64GB LPDDR5x (8 channels). PCIe Gen5: 8 lanes; Gen4: 3 lanes. TDP: estimated 18–45W. Memory bandwidth and PCIe lanes vary by model, with higher-end models offering greater expandability. The N1X (20 cores) features a 16-channel memory interface, achieving the bandwidth inherited from the DGX Spark. On the other hand, the lower-end models use an 8-channel configuration to balance cost and power consumption.

Intensifying Competition with Qualcomm The

Windows on Arm space has so far been largely dominated by Qualcomm. Since the first viable Windows on Arm laptops were introduced in 2018, the Snapdragon series has led the field. However, NVIDIA’s entry into the market may change the landscape significantly. Recently, Qualcomm announced the Snapdragon C series, a low-end Arm chip aimed at laptops starting at $300. In contrast, NVIDIA appears to have no intention of targeting the budget market. Leaked information suggests that NVIDIA’s new chips will primarily target high-end developers and professionals, with significant price differences compared to low-end products. However, NVIDIA holds a significant advantage in GPU performance. The Blackwell 2.0 architecture offers discrete-class graphics performance and Tensor cores optimized for AI inference. This enables support for workloads such as natively running AI applications on Windows on Arm and handling game ray tracing, tasks that were previously challenging for traditional Arm processors. Microsoft’s improved x64 emulation in Windows 11 for Arm also enhances the platform’s usability.

High-End Products for Professionals What is

clear from the leaks is that NVIDIA aims to avoid competing in the low-price segment and instead focus on the premium market. The 20-core N1X outperforms competitors like the Intel Core Ultra 9 and AMD Ryzen AI 9 in terms of CPU core count. Additionally, its integrated GPU, featuring an impressive 6,144 CUDA cores, may match or even surpass the performance of entry-level discrete GPUs such as the GeForce RTX 3050. Given that the DGX Spark was priced at over $3,000, it’s almost certain that laptops equipped with the N1X will exceed $2,000. While specific prices were not listed in the retail data uncovered by WinFuture, NVIDIA is unlikely to lower prices as long as it positions its products as “AI supercomputers.” Meanwhile, the 10-core N1 may be included in more affordable laptops or mini PCs. Yet it is unlikely to compete directly with Qualcomm’s $300 products. NVIDIA’s focus appears to be squarely on high-end users specializing in AI workloads and creative tasks.

Future Outlook The official announcement by

NVIDIA and Microsoft, expected at Computex 2026, will likely reveal specific reference designs and a lineup of partner companies. Laptop manufacturers like ASUS, Lenovo, and MSI are expected to launch products equipped with the N1X series. The Windows on Arm ecosystem has so far relied heavily on Qualcomm. NVIDIA’s entry is expected to stimulate competition and accelerate performance improvements across the Arm platform. Notably, NVIDIA has a clear edge in GPU performance and AI processing capabilities. However, there is also a risk that high prices could limit the market to a niche audience. Compatibility issues with x86 remain a challenge. While Microsoft is working on improving the Prism emulator, the transition barrier remains high unless the number of native Arm applications increases. By expanding the CUDA ecosystem for Arm, NVIDIA could provide developers with powerful tools, which would be key to driving the adoption of Windows on Arm. Although NVIDIA attempted to enter the Windows on Arm market back in 2008, that effort did not materialize. Now, more than 15 years later, the company is taking another shot at this market. By applying the technology developed with DGX Spark to mobile devices, NVIDIA may introduce a new option for laptops in the AI era. The announcement at Computex is eagerly awaited.