Google caps Meta's Gemini use as AI demand strains capacity
Google has restricted Meta's use of its Gemini AI model due to surging demand straining computing capacity, highlighting the intense competition for AI infrastructure resources among tech giants.
Background
- Google has limited Meta’s access to its Gemini AI model (via a cloud API) because the computing infrastructure running it is under heavy strain from surging demand for AI services.
- Gemini is Google’s flagship large language model (LLM), competing with OpenAI’s GPT-4 and Anthropic’s Claude; it powers everything from chatbot apps to enterprise API calls.
- The article draws on a rare public admission by Google that even a major cloud provider cannot keep up with AI compute demand — a sign of how fast AI adoption is outpacing data-center capacity, even for tech giants.
- Meta relied on Google Cloud to run Gemini for certain internal or product uses, and now faces disrupted access; the episode underscores that the AI “gold rush” is creating real supply bottlenecks in chips (Nvidia GPUs), power, and cloud rental capacity.