XLA / OpenXLA
Software
Compute Platform
Active
The compiler that makes TPUs programmable — now an open project.
You're looking at a software record: a tool or platform in the AI stack — who maintains it and what hardware it runs on.
XLA compiles TensorFlow, JAX and PyTorch graphs into optimized code for TPUs and GPUs. Spun into the OpenXLA project with StableHLO as its portable format, it is the main compiler path off the CUDA-only world — every TPU model on this site runs through it.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Supports
Maintained by
Related
Sources
Record reviewed as of 2026-01-15 — see Methodology.