Google overhauls Gemini usage quotas with compute-based limits, squeezing power users

Google overhauled its Gemini AI usage quotas on May 17, replacing fixed prompt-based limits with a compute-based model that measures actual processing power per request, causing power users to hit their limits faster. The change, which includes rolling five-hour cycles and weekly hard caps, sparked backlash leading to adjustments on May 28, including capping single-prompt impact and exempting failed requests. The new system sharpens subscription tier divides, with Ultra users getting 20x higher allowances than free users, and hints at future pay-as-you-go credits that could impact decentralized AI projects.

Google overhauls Gemini usage quotas with compute-based limits, squeezing power users The shift from simple prompt counts to dynamic compute-based caps is already pushing users to hit their limits faster, and it could reshape the AI services landscape. Google just fundamentally changed how it meters AI usage across its Gemini Apps, and the result is that many users are getting fewer responses than they’re used to. The new system, effective as of May 17, ditches the old fixed prompt-based quotas in favor of a compute-based model that accounts for the actual processing power each request demands. In English: instead of counting how many times you ask Gemini a question, Google now measures how hard each question makes its servers work. A simple “what’s the weather” costs less than asking the model to analyze a 50-page PDF. The problem is that power users are burning through their allotments significantly faster under this new framework. How the new system works The compute-based quotas refresh on a rolling five-hour cycle, but there’s a cumulative weekly hard cap that prevents users from simply waiting out cooldowns and binging on heavy prompts. Usage limits now fluctuate based on prompt complexity, the tools invoked during a session, and overall conversation length. This change rolled out shortly after Google I/O 2026 and was designed to better reflect the real computational resources being consumed. Users running the advanced Gemini 3.1 Pro model, which is optimized for complex tasks and large data files, found themselves hitting walls much sooner than expected. The backlash was swift enough that Google issued a round of adjustments around May 28. The company capped the maximum quota impact any single prompt could have, so one particularly gnarly request wouldn’t nuke your entire weekly allowance. Failed requests, which previously still counted against your limits, were exempted from quota consumption. Google also made lighter models like Gemini 3.1 Flash-Lite available at no cost in certain use cases. And perhaps most notably, the company hinted at future pay-as-you-go credit options for AI services. The subscription tier divide The new quota system sharpens the divide between free and paying users. Here’s the breakdown of how limits scale across tiers: Free users get standard limits. Google AI Plus subscribers get 2x. Pro subscribers receive 4x. And Ultra tier users get up to 20x higher allowances compared to the baseline. That’s a tenfold gap between the cheapest paid tier and the most expensive one. What this means for the broader AI and crypto landscape Google hasn’t announced any on-chain billing, token-gated access, or blockchain-based compute metering for Gemini. But the implications for the decentralized AI sector are worth paying attention to. Google’s shift to compute-based billing also validates a pricing model that several crypto-native AI projects already use. Decentralized compute networks like Akash, Render, and others have always priced resources based on actual computational work rather than arbitrary prompt counts. The hint at pay-as-you-go credits is particularly interesting. If Google moves toward a consumption-based billing model for AI services, it starts to resemble the tokenized compute credit systems that several Web3 projects have been building. The difference is that Google’s credits would live on Google’s servers, governed by Google’s rules, subject to Google’s future quota changes. What’s worth watching next is whether Google follows through on those pay-as-you-go credit options and how they’re structured. If the pricing lands in a range that undercuts existing decentralized compute marketplaces, it could be a headwind for Web3 AI projects. If it’s expensive enough to make decentralized alternatives look competitive, it becomes a catalyst. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .