Google Limits Meta's Access to Gemini AI Amid Compute Crunch, Report Says

Meta’s internal artificial intelligence (AI) projects have faced disruptions and delays, and a new report suggests Google’s capacity constraints are partly responsible. According to a Financial Times report published on June 28, Google denied Meta’s request for additional computing capacity in March 2025 and imposed limits on the social media company’s use of its Gemini AI models.

Google informed Meta that it could not fulfil the full Gemini capacity Meta had sought to purchase, citing surging demand for its AI models. The restrictions are not unique to Meta—several other Google customers reliant on the tech giant for compute capacity have also been affected. However, the report notes that Meta has borne the brunt of these limitations.

The development underscores the intense competition for computing resources in the AI sector, where demand for high-performance cloud infrastructure has outpaced supply. Google’s Gemini models are among the most sought-after AI systems, and the company has been allocating capacity to its own AI products and other customers.

Meta has been investing heavily in AI for its platforms, including content recommendation and advertising tools. The interruptions to its AI workstreams could have implications for its product development timelines. Neither Meta nor Google has publicly commented on the report.

The Financial Times report, citing sources familiar with the matter, highlights the broader challenge faced by AI companies: the scarcity of advanced computing power needed to train and deploy large language models. As demand surges, cloud providers like Google, Amazon, and Microsoft are prioritising allocation, sometimes at the expense of existing clients.

Industry analysts note that such capacity crunches are likely to persist as AI adoption accelerates. Companies like Meta may need to diversify their cloud providers or invest in proprietary hardware to reduce dependency on third-party compute capacity.