Google open-sources TorchTPU, a native PyTorch backend for Cloud TPUs

Google has open-sourced TorchTPU 0.2.0, a native PyTorch backend for Cloud TPUs co-developed with Meta's PyTorch team. It replaces PyTorch/XLA, runs via torch.device("tpu") and scales to 8,192 Ironwood chips, with up to 52% MFU reported on Qwen 3 8B training.

• TorchTPU 0.2.0 is on PyPI under Apache 2.0, Python 3.11+
• Reported up to 52% MFU on Qwen 3 8B, scaling to 8,192 Ironwood chips
• torch.compile cut step time 3.8x, from 11.49s to 2.78s
• vLLM, SGLang and Hugging Face Transformers v5 now run on TPU