transformer-patent-google-2019

IN premiseentries/2026/06/21/wiki-Transformer_deep_learning_architecture-chunk-7.md

Created 2026-06-21T09:50:10+00:00

The Transformer is covered by US Patent 10,452,978 ('Attention-based sequence transduction neural networks'), assigned to Google LLC, issued 2019-10-22.

Summary

The Transformer architecture, which underpins essentially all modern large language models, falls under a US patent held by Google, issued in late 2019. This matters because it grounds the system's reasoning about intellectual property: any downstream work, licensing analysis, or legal risk assessment involving transformer-based models can trace back to this specific patent as the root claim.

Dependents

These beliefs depend on this one: