Multiple reports, including the Financial Times and others, say Google has imposed limits on Meta’s access to Google’s Gemini AI models. The move affects how Meta can use Gemini within its products and services that rely on external AI model access. According to these accounts, Google has also asked its employees to focus on using AI tokens more efficiently after the restrictions were put in place. The reports frame the change as part of Google’s tighter management of how its AI infrastructure and model access are used by partners. However, the sources do not provide detailed public terms of the restrictions, such as which specific Gemini model versions are affected, the volume limits (if any), or the duration of the measures. The reports also do not conclusively attribute the decision to a single reason, but they describe it as a practical restriction on model access and usage controls. Overall, the coverage aligns on the core point: Google limits Meta’s use of Gemini models and simultaneously tightens internal guidance around token efficiency.