Google limits Meta's use of Gemini AI models over capacity
Technology
Google has had to cut back Meta's use of Google's Gemini AI models because there just isn't enough computing power to go around.
Meta had been leaning on Google's tech since Gemini was doing better than its own Llama models, but now Google says its servers are maxed out and can't keep up with all the AI requests.
Meta rollout rationed, Google rents SpaceX
This shortage is slowing down Meta's rollout of new AI tools: engineers even had to ration how much they could use.
Meanwhile, Google is juggling a massive $460 billion backlog in cloud contracts and is renting extra computing muscle from SpaceX for $920 million a month.
It's a reminder that even tech giants can run into hardware headaches.