Google Limits Meta’s Gemini Model Usage, Delaying AI Projects

To accelerate internal AI research and development, Meta uses Google’s Gemini models alongside its own models and data centers. As generative AI services expand rapidly, cloud providers face shortages of chips and data center capacity. The supply of computing resources between companies is therefore having a direct impact on product development timelines and competitive strategies.

As of July 20, 2026, reports said Google had capped Meta’s usage of Gemini models, delaying some of Meta’s internal AI projects. Google was also seeking to rent external computing capacity to ease the bottleneck. The reports did not disclose when the restrictions took effect, the size of the reduction, the rental cost or the number of affected projects.