Google ups Gemini limits once again for Antigravity users

Two days ago, we reported that Google quietly nerfed the AI Pro plan, and the backlash was immediate. Over on Reddit, frustrated users accused the company of pulling a bait-and-switch after noticing tighter Gemini usage limits than before. For users who depend on Gemini for coding, deep research, or hours-long workflows, hitting weekly limits so quickly made the AI Pro plan feel unnecessarily restrictive, especially for something people are actively paying for.

Now, Google appears to be scrambling to calm things down. Shortly after the criticism picked up steam, Varun Mohan, Director at Google DeepMind working on Antigravity, posted on X saying the company was tripling Gemini rate limits across all paid Antigravity tiers and resetting everyone’s weekly quota. According to Mohan, the team realized some users were exhausting their limits far too quickly and wanted to respond fast. But that wasn’t the end of it.

In another update posted today, Mohan announced that Google is again tripling weekly Gemini quotas across all paid plans in Antigravity and even resetting them.

If you’re keeping count, that effectively works out to a massive 9x increase compared to where limits landed after the original nerf. And honestly, that kind of rapid reversal says a lot.

This means people are using Antigravity far more aggressively than Google expected. Mohan said the company saw a huge jump in users building projects after the first quota increase, while many others worried they’d burn through their weekly allowance after just a couple of work sessions.

That said, this doesn’t magically erase the bigger issue. The increased limits only apply to Antigravity right now, not Gemini as a whole. Outside that environment, usage caps remain unchanged, which means plenty of paid users will still feel squeezed elsewhere in Google’s AI ecosystem. And users were quick to point that out. Even with the latest increase, several people noted that limits are still effectively lower than what existed before Google tightened things in the first place. So while the company is clearly trying to course-correct, this feels more like damage control than generosity.

The bigger takeaway here is that Google may have underestimated how quickly people are integrating Gemini into real workflows. Once users start depending on AI tools for hours-long coding sessions, research dumps, or project planning, aggressive rate limits start becoming productivity blockers. That’s especially risky at a time when competitors are pushing harder on unlimited or at least more flexible AI access.

Right now, Google seems caught between two goals: keeping infrastructure costs under control while also convincing users that its premium AI plans are worth paying for. Judging by how quickly the company reversed course twice in less than 48 hours, it’s pretty clear which side users are forcing Google toward.