On one hand, Google is making headlines regarding its I/O 2026 event; on the other, however, the company is facing massive backlash from users. As it turns out, Google has quietly imposed a usage limit on its subscription plans. Let’s take a detailed look at exactly what the entire matter entails.
At its I/O 2026 developer conference held on Monday, Google unveiled several major AI updates, including the Gemini Spark Agent Platform, AntiGravity 2.0, the Gemini Omni World Model, and Gemini 3.5 Flash, among others. However, on the flip side, Google has managed to upset many of its users. The issue revolves around Google’s AI plans, which the company has quietly reshuffled. According to a report, the company has silently implemented changes to its Google AI plans; specifically, the top-tier AI Ultra plan is now available with a $50 discount, starting at $200, while a new Google AI Ultra tier has been introduced at $100 per month. Google has also quietly made another change: reducing the usage limits for its AI Pro subscription—a service that, incidentally, is currently available for free to many Indian users holding a Jio SIM card. These new changes to the Google AI Pro plan are causing frustration among social media users, with netizens venting their anger on platforms like Reddit.
Google Quietly Changes Terms of AI Pro Plan
According to a report by Android Authority, Google has begun notifying AI Pro subscribers via email about the changes being implemented in the plan. It appears that Google has introduced a new five-hour rolling window system, similar to the approach taken by Anthropic. This means that once users exhaust their compute quota, they must wait for a five-hour cycle to complete before their quota refreshes.
Google stated: “For the Gemini app, we are introducing compute-based usage limits that take into account the complexity of your prompts, the features you utilize, and the length of your chats. Your limits will refresh every five hours, provided you have not yet reached your weekly limit.” According to Android Authority, an email sent by Google states that AI Pro subscribers will receive a usage limit “four times higher” than that of free users; however, Google did not specify the exact count of tokens or prompts.
Google Quietly Reduces Usage Limits
Internet Users Frustrated by Changes to Google’s AI Pro Plan
Immediately after the changes to Google’s AI Pro plan went into effect, complaints from disgruntled users began surfacing on Reddit—many of whom even threatened to cancel their subscriptions in protest.
One user wrote, “They didn’t even bother sending an email regarding this major contractual change.” They continued, “Today, I burned through 50% of my entire 5-hour limit in a single response to just five posts. I’m not going to use AI like an energy meter in a mobile game, and I certainly won’t start a new chat every five minutes just to avoid the heavy ‘compute tax’ imposed on long threads.”
Another subscriber expressed astonishment at how quickly their limits were depleted: “My 5-hour quota ran out after exactly four prompts. These tasks involved generating a PDF, creating an image, and explaining code—all executed using the Pro model. During the subsequent quota cycle (after a reset), even a failed generation attempt was counted against the limit (specifically, I simply received a message stating that the model was under heavy load and to try again later). This occurred multiple times, and each instance was counted toward the quota.”
Many other users also expressed frustration with the five-hour usage limit.
One user wrote, “At the very least, they could have set a daily quota, man. Why specifically every five hours?” They added, “I don’t use Gemini while I’m sleeping or during certain parts of the day… It feels like they’re trying to mimic Anthropic and think they can get away with it.”
Other users on Reddit also complained that even when media generation attempts fail, they are still counted toward the AI Pro usage quota. Some users even claimed that despite selecting “Pro” in their model preferences, they were switched from the Pro model to the Flash version.
