Who contributed to Google on August 17, 2026?

4 developers shipped this update, including copybara-service, Benjamin Chetioui, jax authors, and zhixiangli.

What were the notable Google updates?

[Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs, [Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs, and Update XLA dependency to use revision.

Google

JAX, the GenAI SDK, and the Cloud libs - Google's open source layer

Pick a date

Topics: AI / ML Python Full archive →

The Wire · Showcase

JAX ROLLS BACK TPU ACCUMULATOR API, CLOUD STORAGE FIXES STREAM CLEANUP

By RepoJournal · Filed 07:06 UTC on August 17, 2026 · About Google

4 people shipped this

zhixiangli @zhixiangli 2 cited

copybara-service @copybara-service 1 cited

Benjamin Chetioui @bchetioui 1 cited

jax authors @Google-ML-Automation 1 cited

JAX reverted a Pallas TPU accumulator change after it broke the explicit MXU API surface, while Google Cloud Storage patched a critical gRPC stream cleanup bug that could leak resources on connection failures.

The JAX team rolled back an accumulator reference exposure [1] [2] that was causing issues with downstream TPU workloads relying on explicit MXU APIs, restoring stability across the Pallas subsystem. In parallel, the XLA dependency received an automated update [3] to keep the compiler in sync. Over in Cloud Storage, a fix landed that ensures bidirectional gRPC streams properly close on open failures or cancellations [4] [5], addressing a resource leak in _AsyncReadObjectStream that could strand sockets if initialization failed. The storage fix adds a dedicated _close_socket_like_rpc helper and guards requests_done() calls to prevent double-closure failures, backed by new unit tests for the failure paths.

Action items

→ Update google-cloud-storage if you use async blob reads, especially in retry-heavy workloads googleapis/google-cloud-python [plan]
→ If running JAX on TPU with explicit MXU calls, verify your workload still passes after the accumulator rollback google/jax [monitor]

References

[1] [Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs. ↗ google/jax
[2] [Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs. google/jax
[3] Update XLA dependency to use revision google/jax
[4] fix(storage): ensure bidi-gRPC stream is cleaned up on open failure and close ↗ googleapis/google-cloud-python
[5] fix(storage): ensure bidi-gRPC stream is cleaned up on open failure and close (#18119) googleapis/google-cloud-python

Quick answers

What shipped in Google on August 17, 2026?: JAX reverted a Pallas TPU accumulator change after it broke the explicit MXU API surface, while Google Cloud Storage patched a critical gRPC stream cleanup bug that could leak resources on connection failures. In total, 3 commits and 2 pull requests landed.
Who contributed to Google on August 17, 2026?: 4 developers shipped this update, including copybara-service, Benjamin Chetioui, jax authors, and zhixiangli.
What were the notable Google updates?: [Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs, [Pallas TPU] Roll back exposing accumulator references to be used with explicit MXU APIs, and Update XLA dependency to use revision.

JAX 0.11.1 lands with export guardrails and a tracing speedup

JAX's latest release locks the door on expired exports while a core tracing hot path gets a serious speed boost.

ai 63 shipped

Google Aug 16

JAX EXPANDS MPMD DEBUGGING, GOOGLE CLOUD PYTHON PLUGS BIGTABLE DATA LOSS

JAX's callback system now works across multi-machine distributed setups, while Google Cloud Python fixes a critical bug that could silently drop database mutations.

ai 17 shipped

Google Aug 15

JAX SHIPS NATIVE TPU CONVOLUTION SUPPORT, PYTHON 3.15 CONSTRAINT LIFTED

JAX Pallas now has native convolution operators for TPU, part one of a three-part overhaul that unlocks spatial N-D convolutions directly in the MLIR Mosaic dialect.

ai 62 shipped

Google Aug 14

JAX LOCKS IN PYTHON 3.15 SUPPORT AS GENAI SDK SHIPS VIDEO ENHANCEMENTS

JAX is ready for Python 3.15 and its free-threaded runtime, while Google's Genai SDK adds video resolution controls and BigQuery gets Apache Arrow query results.