Google releases TorchTPU, letting PyTorch apps run on TPUs without rewriting

Google announced the open-source release of TorchTPU, its long-developed PyTorch-adapted version that lets AI apps run on Google's in-house TPU chips. Existing PyTorch apps can move to TPUs and Google Cloud by changing configuration alone, with no code rewrite. Google says it worked directly with Meta, Hugging Face, Inferact (vLLM) and RadixArk (SGLang) to keep TorchTPU's codebase compatible with these mainstream tools. It also launched a TPU agent suite to help developers migrate apps and fix errors, and updated LiteRT, its tool for running AI on phones, laptops and embedded devices, to better support PyTorch apps. For developers and startups building and deploying models with PyTorch, the barrier to moving to TPUs drops from rewriting code to editing configuration, adding a hardware and cloud option beyond Nvidia GPUs. The original notes, however, that this does not mean developers will abandon Nvidia hardware at scale anytime soon.