Google has joined Amazon in attacking Nvidia’s stronghold with AI app developers
Google released TorchTPU, an open-source PyTorch version for TPUs, challenging Nvidia’s dominance, after Amazon’s similar TorchNeuron announcement.Summary
On Tuesday, Google announced TorchTPU, an open-source version of PyTorch that runs AI applications on Google’s Tensor Processing Units (TPUs) and Google Cloud. This follows Amazon’s Monday release of TorchNeuron, which ports PyTorch apps to Amazon’s Trainium chips and AWS. Both moves aim to reduce Nvidia’s stronghold on the AI industry, where most AI apps are built for Nvidia GPUs. PyTorch, developed by Meta in 2016 and now managed by the Linux Foundation, is used by 85% of AI apps. Google says it worked with Meta, Hugging Face, and others to ensure TorchTPU compatibility, and launched a TPU agent suite to help developers migrate and debug. Amazon is also contributing code to PyTorch to decouple features from Nvidia hardware. While developers may not immediately abandon Nvidia, Google and Amazon are clearly working to erode its advantage.
(Source：TechCrunch)