Huawei Ascend chips now support the Triton language, enabling developers to write high-performance GPU kernels without CUDA. This move breaks the Nvidia software monopoly by simplifying hardware-specific optimization. Practitioners can now port AI workloads to Ascend hardware more efficiently. It reduces the technical barrier for deploying large models on non-Nvidia silicon.