Explained: Why Google moved Gemini to token-based limits

Google has tightened how Gemini's free and paid tiers work since May 17, 2026, introducing compute-based usage limits that refresh every five hours up to a weekly cap. The change is driven by a compute shortage that has affected Google's largest enterprise customers for months. The limits depend on prompt complexity, model chosen, and chat length, effectively rationing finite compute resources across the user base.

Meta was the first major customer to be hit. In March, Google capped Meta's use of Gemini models after Meta demanded more capacity than Google could supply, as reported by the Financial Times. The restriction disrupted some of Meta's internal AI projects and forced the company to ask staff to use AI tokens more efficiently. Other Google clients faced similar but less severe constraints, highlighting the seriousness of the shortage.

Google CEO Sundar Pichai admitted the company is "compute-constrained in the near term" during the first-quarter earnings call in April. Cloud revenue exceeded $20 billion for the first time, while signed-but-undelivered cloud contracts nearly doubled to over $460 billion. This backlog indicates that customers are signing up for compute faster than Google can deliver, underscoring the structural nature of the shortage.

To cope, Google signed a $920 million-per-month deal to lease computing capacity from Elon Musk's SpaceX, and Anthropic made a similar arrangement. These leasing deals suggest the shortage is a long-term market feature. The highest cost now comes from inference—the work of running models after training—meaning every user prompt consumes compute, making everyday AI usage increasingly expensive to support.

Google's help page explicitly ties limits to capacity: it may cut free user limits before paid subscribers during capacity changes, withhold compute-heavy features like Deep Research during high demand, and adjust limits without notice. Token-based limits allow Google to charge heavier tasks more and steer demand toward lighter models. For India, where Google built its user base on free access, these limits mark a shift from a "scale first, monetise later" strategy to actual monetisation, as seen in free giveaways like the Google-Jio 18-month AI Pro plan.