Google Caps Meta's Gemini Access as AI Demand Strains Capacity
The restrictions reveal that even the largest AI companies face infrastructure bottlenecks, with demand from a single enterprise customer like Meta directly straining Google's computing capacity.
Reporting from 1 sources: GIGAZINE.
Google has imposed restrictions on Meta's access to its Gemini AI models, unable to provide the computing capacity Meta requested. The disruption began around March 2026 and was still ongoing as of late June. Some of Meta's internal AI projects have been delayed, and Meta employees are being encouraged to use AI tokens more efficiently. Other Google customers are also affected, but not to the same extent as Meta.
Google has restricted Meta's access to its Gemini AI models, according to sources familiar with the matter. The computing capacity Google agreed to provide to Meta has not been delivered, forcing Meta to cap its usage. The disruption, which started around March 2026, was still in effect as of the end of June. Some of Meta's internal AI projects that depended on full Gemini capacity have been delayed. Meta employees are now being told to use AI tokens more efficiently, partly due to these restrictions and partly due to Meta's own cost-reduction efforts.
One source said other Google customers are also affected by the restrictions, but not to the same extent as Meta, whose demand for Google's models is particularly large. Google's cloud business revenue exceeded $20 billion in its April 2026 earnings report, and its backlog of signed but undelivered cloud contracts nearly doubled quarter-over-quarter to over $460 billion. However, Google is struggling to secure infrastructure. In June 2026, Google signed a contract with SpaceX to access computing capacity. Meta is also partnering with AMD and NVIDIA for its own AI strategy.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.