After Amazon, Google targets Nvidia’s software lock-in with TorchTPU
Google open-sourced TorchTPU, enabling developers to run AI workloads on its Tensor Processing Units (TPUs) with minimal code changes. This follows Amazon's launch of TorchNeuron, aimed at reducing reliance on Nvidia's software ecosystem. The release, announced at the PyTorch Conference 2026 on October 20, enhances flexibility for developers in hardware selection.