A way to pause some paid services
Google made spending caps in Cloud Budgets generally available on October 30. They are meant for people running paid services on Google Cloud, rather than ordinary Gemini chat users. The update adds management through an API, a software connection that lets programs work with the caps, along with faster cost estimates and automatic monthly resets. Google says caps on AI services can trigger within minutes. [1]
What can still cost money
Each monthly cap covers one project and one eligible service: Gemini API, Gemini Enterprise Agent Platform, Cloud Run, or Cloud Run functions. Estimates leave out discounts and credits. Subscriptions and reseller accounts are excluded. [2]
Reaching a cap pauses new usage without deleting your resources. Requests already running can finish, fixed costs such as storage can continue, and you still owe charges above the cap. Reserved computing capacity called Provisioned Throughput keeps running. A monthly reset can automatically restart paused services. We have not tested enforcement or billing. [1][2]
Our take
If you're experimenting with an AI app, list everything it pays for before relying on a cap. Leave room below your actual spending ceiling for charges that arrive late or keep running. Plan for the app to pause, and put the next month's automatic restart on your calendar.