The 5 October launch arrives alongside two separate public TPU workload disclosures from Google Cloud this week.
Google announced the release of Gemini Omni 1.1 Flash, an AI model for video generation and editing, on 5 October 2026.
This model launch landed in the middle of three consecutive public disclosures from Google relating to TPU deployment for generative AI workloads.
On 4 October, Google published confirmation that video generation vendor HeyGen had ported its Avatar IV model to Google Cloud Trillium v6e TPUs. That port was originally announced by HeyGen on 13 August 2026.
One day prior, Google staff software engineer Ran Ran published that a Google Cloud team had reproduced the Allen Institute for AI's Olmo 3 7B language model from scratch, running the build on Google Cloud TPUs using the MaxText framework.
No utilization rates, contract terms, or capex figures were included with any of the three announcements.
Read with its sources →