Google has capped Meta's access to its Gemini AI models after the social media giant requested more computing capacity than the tech company could fulfill, according to a Financial Times report. The restriction emerged around March when Google informed Meta it couldn't meet the full demand for Gemini model capacity. This supply constraint has disrupted several of Meta's internal AI projects, forcing delays in development timelines.

Other Google Cloud customers have experienced similar limitations, though Meta faces the most severe impact due to its exceptionally high demand for AI computing resources. In response, Meta has reportedly encouraged staff to use AI tokens more efficiently, treating these computational units as precious resources amid the shortage. The incident reveals intensifying competition for AI infrastructure among tech giants.

Despite companies investing billions in chips and data centers, computing power remains insufficient for surging AI service demand. Google Cloud revenue reached $20 billion in the first quarter ending March, yet CEO Sundar Pichai acknowledged that compute constraints prevented even stronger growth. The cloud unit's backlog nearly doubled sequentially, reflecting pent-up demand across Alphabet's client base.

This capacity crunch highlights the bottleneck facing AI development at scale. While Meta continues advancing its own Llama models, reliance on external providers like Google creates vulnerabilities when supply falls short of ambitious AI expansion plans. The situation underscores infrastructure scarcity as a critical challenge for the industry.