Google launches TorchTPU to loosen Nvidia's software lock-in

What happened

Google has released TorchTPU as an open-source project, allowing PyTorch applications to run on its tensor processing units with minimal configuration changes. It also launched a TPU agent suite to help developers migrate workloads and fix errors.

Why it matters

Nvidia's advantage comes not only from chip performance but also from the software ecosystem built around its hardware. Easier PyTorch portability lowers one of the practical barriers preventing developers from testing competing accelerators.

The bigger picture

Cloud providers are increasingly competing at the developer-tooling layer as they try to move AI workloads onto their own chips.