What happened
Google restricted Meta's access to its Gemini AI models around March, citing insufficient computing capacity. This limitation disrupted and delayed some of Meta's internal AI projects, with other Google clients also experiencing lesser impacts. Meta's exceptionally high demand for Gemini models led to the specific restrictions, prompting the company to encourage staff to optimise AI token usage. Google Cloud generated $20 billion revenue in the first quarter, but CEO Sundar Pichai stated computing power constraints prevented higher growth and nearly doubled the cloud unit's backlog.
Why it matters
Restricted access to frontier AI models will constrain large-scale AI development for platform engineers and founders relying on external providers. This capacity limitation, driven by high demand and Google's own internal needs, directly impacts project timelines and resource allocation, forcing companies like Meta to optimise existing AI token usage. The constraint on compute power, despite significant investment in data centres, indicates a persistent bottleneck for scaling AI initiatives, affecting procurement teams planning future model access.
