Meta deploys Google TPUs in production for AI

Meta has deployed Google’s tensor processing units (TPUs), its custom AI chips, in production to power AI workloads, according to a new announcement. This marks the first time Google has sold TPUs to external customers, challenging Nvidia’s market dominance. The new versions of the SGLang and vLLM inference engines now support PyTorch natively, allowing Meta to use TPUs without major rewrites of its code. This development comes after a close collaboration between Meta, Google, and the SGLang team, and it enables Meta to deploy PyTorch workloads on Google TPUs. This follows Meta’s earlier attempts to use TPUs through an older bridge from PyTorch to TPUs.

Previously, both engines used Google’s JAX AI framework to run on TPUs. The new versions allow the hardware to be a native option for PyTorch, mostly without rewriting. This is not Meta’s first attempt at using TPUs. Google calls PyTorch/XLA, the older bridge from PyTorch to TPUs, a 'predecessor' to the new TorchTPU, and SemiAnalysis reported last year that Meta research teams used TPUs through this bridge from 2020 to 2023 before Meta canceled those contracts. Google’s PyTorch-on-TPU push was aimed mainly at Meta, which did not want to migrate to JAX.

Although it is unclear where Meta’s production TPUs sit, talks last year reportedly included Google Cloud TPU rentals in 2026 with purchases in 2027, and Meta later reportedly signed a multi-year rental agreement. This announcement came one day after Amazon said it was open-sourcing TorchNeuron, its own native PyTorch backend for its Trainium2 and Trainium3 chips, at the same PyTorch Conference. Both moves appear to be aimed at loosening Nvidia’s software lock-in.

On performance, a tuned vLLM setup serving Moonshot AI’s open-weight, 2.8-trillion-parameter Kimi K3 model reportedly tops 700 tokens per second for a single user on the seventh-generation Ironwood TPU. That speed figure comes from Inferact, the startup founded by vLLM’s creators that helped Google build vLLM-TorchTPU. Its test ran 16 Ironwood chips with speculative decoding, a method of speeding up large language model inference with a small draft model, against vLLM’s published setup on 16 Nvidia GB200s, which reached 452 tokens per second. Its kernels for Kimi K3 and Qwen 3.8 27B deliver roughly 1.4 to 2 times the GB200 baseline’s throughput at small batch sizes without speculative decoding.

SGLang-TorchTPU will keep adding models, starting with DeepSeek V4.1 Flash and Reflection Beam. It's also working toward Day 0 TPU support for new open models, meaning support on release day, by the first quarter of 2027, according to RadixArk. Google, Meta and RadixArk plan to share more at SGLang Summit 2026, Nov. 12-13 in San Francisco.