ExecuTorch Becomes a Part of PyTorch Core to Expand On-Device Inference Capabilities – PyTorch

ExecuTorch, Meta's runtime for running PyTorch models on edge devices, is joining PyTorch Core under the PyTorch Foundation. This move provides vendor-neutral governance, clear IP and trademark stewardship, and a transparent multi-stakeholder structure. Originally built at Meta for mobile, AR/VR, and embedded deployments, ExecuTorch has grown into a broader ecosystem used for generative AI, LLM assistants, computer vision, and low-latency on-device applications. The transition formalizes technical governance aligned with PyTorch's existing maintainer hierarchy and sets priorities around expanding hardware coverage, deepening PyTorch integration, and growing the contributor base.