Google told Meta it cannot supply as much Gemini compute capacity as Meta wants. Meta had been using Gemini for content moderation at massive scale. Now Meta is shifting that work to its own Muse Spark model to reduce dependence on Google.
This is the sound of a constraint becoming visible. Google Cloud quarterly revenue topped $20 billion for the first time in March. The backlog of signed but undelivered contracts roughly doubled to $460 billion. That is not revenue they earned. That is revenue they committed to but cannot yet deliver.
So Google went to SpaceX and bought roughly 110,000 Nvidia GPUs housed in xAI data centers at $920 million a month just to bridge the gap. A billion dollars a month to temporarily rent someone else’s compute while they build more capacity. That is not a market signal that demand will cool. That is a market signal that even the largest cloud provider cannot keep up.
And Meta, one of the largest AI consumers, is now forced to fall back on its own models rather than depend on someone else’s. You move to your own silicon when you cannot trust the supply chain to deliver at the scale you need.
This is how the industry fractures. Not through hostility or force, but through sheer scarcity. Google cannot supply. Meta cannot risk depending on Google. Microsoft is betting Anthropic will use Maia chips. Amazon is betting on custom silicon too. Everybody is building their own models, their own chips, their own supply chains. The centralizing moment of three companies running the frontier has passed.
——
Follow: @Ali Demi
Book your free AI clarity call, NOW!
https://buff.ly/TpWy277
——
Sources:
https://www.forbes.com/sites/jonmarkman/2026/06/29/google-limits-metas-gemini-usage-over-compute-shortages/
https://thenextweb.com/news/google-caps-meta-gemini-compute-shortage
https://qz.com/google-meta-gemini-ai-compute-shortage-062926
Repost this. Thanks.
