In March of this year, Google informed Meta that it would not be able to provide the requested volume of access to its Gemini artificial intelligence platform. This was reported by the Financial Times.
According to the report, while Meta has its own AI models, they are considered weaker compared to solutions from rivals like OpenAI, Claude, and Google. For this reason, Meta had been using Gemini models to improve its fraud and harmful content detection systems, optimize customer services, and increase the efficiency of internal business processes and programming.
The Financial Times writes that Google's decision has caused delays and technical difficulties in the implementation of some of Meta's internal AI projects. At the same time, due to the high financial burden of using AI technologies, the company has urged its employees to use allocated AI tokens more economically and efficiently.
Additionally, Meta has begun a phased transition of its employees to the company's new "Meta Spark" model. This model is reportedly more competitive with Gemini in terms of capabilities and serves to reduce Meta's dependence on Google's technologies.