Capacity Constraints Hit Google Cloud
Google has placed restrictions on Meta’s access to its Gemini AI models, as surging demand for artificial intelligence computing power outstrips available infrastructure. Sources familiar with the situation said Google could not fulfill the computing capacity Meta requested under their existing agreement, prompting usage caps that began around March 2026 and remain in effect as of late June. The shortfall has delayed several of Meta’s AI projects that depended on full access to Gemini’s processing power. Meta has responded by urging employees to use AI tokens more efficiently and is pursuing broader cost cutting measures.
Widespread Impact and Growing Backlog
Other Google Cloud customers have also experienced restrictions, though none as severe as Meta, which has unusually high demand for the models. The bottlenecks underscore infrastructure challenges facing even the largest AI companies. In its April 2026 earnings, Google reported cloud revenue exceeding $20 billion for the first time, while its backlog of signed but undelivered cloud contracts nearly doubled quarter over quarter to over $460 billion. To address the shortfall, Google contracted with SpaceX in June 2026 to access additional computing power, reportedly paying nearly $1 billion per month.
Source: Financial Times
