Google Gemini's new usage limits are already frustrating subscribers [U: Limits increased]

Update 21/05/26 – 12:49 pm (IST): Well, that was quick. Just a day after the widespread backlash against the new weekly usage limits, Google has confirmed that the company is resetting the usage limit for everyone this week. 

Additionally, the company is now boosting usage limits for all paid tiers by 3x, permanently. So this isn’t a high-usage period promotion for a limited time just to silence the angry crowd.

Original article published on May 20, 2026, follows:

Google has quietly rolled out strict new usage limits for its paid Gemini AI tiers following the I/O 2026 event, and subscribers are already feeling the pinch.

Paying customers on Gemini Pro and Ultra plans report that a handful of prompts or a few video generations are now entirely depleting their quotas. It is locking them out of the service for up to five hours.

This compute crackdown is just the latest headache for Google’s developer ecosystem today. It follows a widespread ‘Error 253’ bug that completely locks paying subscribers out of Google Flow and AI Studio. It also arrives right alongside a messy Antigravity 2.0 update that stripped out the platform’s built-in code editor and broke existing workspaces.

One frustrated user posted on Reddit that they canceled their Plus subscription after a massive, unannounced contract change. They said a simple five-post back-and-forth burned through 50 percent of their entire five-hour compute limit.

They refused to treat the AI like a mobile game energy meter. Other users in the same thread noted that asking Gemini to re-read chat context now costs a heavy premium. The post now has over 750 upvotes and over 200 comments in under 24 hours.

Similarly, one user in another thread said that they are canceling their Gemini Pro subscription. They noted that they hit their limit after just five prompts in a day. They had previously used the AI for coding with almost no interruptions for months.

Now the leash is significantly tighter. The user advised others to cancel their plans as well.

Over on X, Leon Lin pointed out how fast the new Gemini Omni model drains allowances. Generating just five videos on the expensive Ultra tier wiped out their quota completely. That translates to roughly 17 to 19 percent of the limit per single video generation.

The restrictions appear tied to a new compute model that aggressively punishes long chat histories and complex tasks. Developer Uves Arshad posted that he used to run Gemini Flash for hours straight. Now he drops below 25 percent capacity in just 30 minutes.

One user on the Google Support forums noted that asking the AI to summarize a single document consumed over 25 percent of their usage.

Another discovered that having the personalization feature turned on drastically shrinks the available limit.

Google has not clearly communicated these backend changes to users. People are simply logging in, running their normal daily workflows, and suddenly hitting an invisible ceiling.

That said, Google isn’t the only one pulling back the reins either. Out of nowhere, xAI also seemingly cut down usage limits drastically for Grok subscribers. Following the backlash, Elon Musk claimed that Grok usage limits will be increased, but offered absolutely no timeline for the adjustment, leaving paying users on both platforms stranded in the meantime.

For those interested, here are more reports for reference: 1,2,3.

We stand out from the tech-media crowd because we break news stories; we mainly bring you stuff that you won’t find anywhere in the mainstream tech media. Our stories have been picked up by some of the world’s most popular websites and media outlets—more info is available here.