CUDA 13.4 ready for RTX Spark

Ava Taylor 09/11/2026

NVIDIA Unlocks New Frontiers: CUDA 13.4.1 Brings Windows on Arm Support to RTX Spark

The world of GPU computing just got a significant upgrade. NVIDIA has officially rolled out the CUDA Toolkit 13.4.1, making CUDA 13.4 generally available to developers and enthusiasts. But what makes this release truly noteworthy is the seamless integration of new platform support, specifically bringing crucial Windows on Arm capabilities to the increasingly popular RTX Spark series of graphics cards.

This development is a major step toward democratizing high-performance computing. By extending support to the Windows on Arm environment, NVIDIA is effectively bridging the gap between cutting-edge AI and machine learning applications and a wider range of accessible hardware platforms. This move opens up exciting new avenues for users and developers who are looking to leverage the power of NVIDIA technology on diverse and modern systems.

The immediate impact is felt most strongly by users of the RTX Spark lineup. These cards, designed for versatility and performance, now benefit from enhanced compatibility and optimized performance through the updated CUDA framework. This means that the creative potential unlocked by these GPUs is no longer confined to traditional setups; it can now thrive in a more flexible, ARM-based ecosystem.

For software developers, this release signifies a clear direction. Supporting Windows on Arm significantly reduces barriers to entry, allowing new projects and applications to be developed and tested on a broader range of devices. It fosters an environment where innovative ideas can be built without being constrained by proprietary hardware limitations.

The combination of powerful GPU capabilities and flexible operating system support positions the RTX Spark and associated hardware at the forefront of mobile and embedded computing. NVIDIA’s continued focus on pushing the boundaries of what is possible with CUDA ensures that the push for innovation remains relentless, making complex graphical and computational tasks more accessible than ever before.