Google Limits Meta's Access to Gemini AI Amid Tech Giants' Computing Power Crunch

Google has restricted Meta's usage of its Gemini AI model after Meta requested more computing capacity than Google could provide, according to the Financial Times. The AI revolution has sparked a multibillion-dollar race among major tech companies to invest heavily in computing infrastructure to support advanced AI model training and deployment. This surge in demand has caused increased consumption of processors and memory chips, leading to shortages and significant price hikes in the consumer market as companies strive to build more powerful data centers.

Despite these efforts, demand continues to outpace supply, and even tech giants are now feeling the strain. Sources told the Financial Times that in March, Google informed Meta it could not fulfill the full Gemini capacity Meta sought. Although Meta has its own AI models, they are considered weaker than competitors like OpenAI, Claude, and Google’s models. Meta uses Google’s Gemini models to enhance fraud detection, harmful content identification, customer service, and internal workflows such as programming and efficiency improvements.

Google's decision to limit Meta's access has disrupted or delayed some of Meta's internal AI projects. Due to the high costs associated with AI usage, Meta is encouraging employees to use AI tokens more efficiently. Concurrently, Meta is transitioning staff to its new AI model, Meta Spark, which is seen as a stronger competitor to Gemini and helps reduce Meta’s reliance on Google. Google has also limited usage for other clients, but Meta is the most affected due to its dependence on Gemini. Both companies declined to comment on the report.