That sinking feeling when your API bill rockets past your budget overnight just got a 10-minute grace period Google’s new kill-switch arrives, but the meter keeps running.
Google AI Studio now lets any Gemini API user cap monthly spend per project. Set the dollar ceiling once; it stays locked until you change it. The catch: enforcement lags roughly ten minutes, and every token fired in that window still hits your card.
Quote from the dashboard fine print:
Spend caps have a ~10 minute delay and users are responsible for overages incurred during that period.
For a lean startup, the math is brutal. A $500 cap can absorb a $20 burst if a promo video goes viral and your chatbot spikes from 50 to 5 000 queries per minute. You stop the burn only after the damage is done.
The upgrade pipeline has also shifted from manual pleas to autopilot. Usage Tiers now climb automatically when your combined spend and payment history clear lower thresholds.
All of this surfaces inside a redesigned console. A single page tracks Requests Per Minute, Tokens Per Minute, and Requests Per Day; a daily cost graph can be sliced by model and window (7–30 days). No more tab-hopping to Google Cloud Console to learn which model ate the budget.
The controls are live today at no extra cost; standard API rates still apply.
Source: Google