Users of the paid Google AI Pro subscription encountered an unexpected problem: a single video generation request completely exhausted the five-hour Gemini usage limit. The incident sparked widespread backlash and forced Google to respond.
New quota system: what went wrong?
Google recently switched to a new quota calculation model for Gemini. Instead of a fixed number of requests, a "computational" system is now applied: consumption depends on task complexity, dialogue length, and the functions used. Limits refresh every five hours, but a weekly volume is also taken into account.
One user's story
User Ashutosh Shrivastava posted a screen recording on X showing that after four minutes of processing a video generation request, the system displayed 100% limit usage — even though the video itself was never created. This triggered a wave of outrage among subscribers.
Google's reaction
Gemini lead Josh Woodward publicly acknowledged the issue and promised to investigate. However, this did not calm users, who are mass-reporting a sharp reduction in available quota and the opacity of the new model.
Why does this matter?
Criticism is linked not only to restrictions but also to a lack of predictability. Users cannot understand in advance how many resources a specific task will "consume." This creates inconvenience and reduces trust in the service.