In March of this year, Google significantly restricted Meta's access to its Gemini models. The reason was a lack of computing power to handle the full volume of requests from Meta. This led to the disruption and slowdown of several of the company's internal AI projects, and the restrictions have not yet been lifted.

Against the backdrop of these issues and a general drive to cut costs, Meta urged its employees to use tokens more sparingly. Other Google clients also faced limits, but to a much lesser extent.

This situation highlights the vulnerability of major technology corporations that depend on external cloud providers. Meta, which is actively investing in developing its own AI models, found itself in a position where critically important projects depend on a competitor's resources.

My analysis: This incident is a clear signal that the cloud computing market for AI is becoming a field of fierce competition rather than partnership. Meta will likely accelerate the construction of its own data centers and the development of more efficient architectures to reduce its dependence on Google. For investors, this means rising capital expenditures for major players and increased vertical integration in the AI sector.