GoKawiilGoogle releases TorchTPU, an open-source PyTorch version for its own AI chips
Google announced TorchTPU, a version of the PyTorch framework built to run on its Tensor Processing Units, letting developers shift existing PyTorch-based AI apps to Google Cloud with a simple configuration change rather than a rewrite. The move follows Amazon's release a day earlier of TorchNeuron, its own PyTorch variant for AWS's Trainium chips, and Amazon's submission of code to decouple PyTorch features from Nvidia hardware.
GoKawiil's interpretation of the reporting above, not reported fact.
PyTorch is reportedly used by around 85% of AI applications and was historically optimized for Nvidia GPUs, so making it run natively on custom chips could reduce developer lock-in to Nvidia hardware. Google's collaboration with Meta, Hugging Face and other tool makers suggests an attempt to ensure TorchTPU integrates as smoothly as the original framework, which could make switching to Google's cloud infrastructure more appealing to developers weighing alternatives to Nvidia-based systems.
- Google released TorchTPU, an open-source PyTorch version for its Tensor Processing Units.
- Amazon released a similar tool, TorchNeuron, for its Trainium chips a day earlier.
- Both moves aim to let AI apps run on non-Nvidia hardware without rewriting code.
Source: techcrunch.com — Julie Bort, 2026-10-20
Published there as: “Google has joined Amazon in attacking Nvidia’s stronghold with AI app developers”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.