Google told Meta around March that it could not supply the full Gemini AI capacity Meta wanted to buy, a shortfall that disrupted and delayed some of Meta’s internal AI projects, according to the Financial Times. Several other Google clients were affected to a lesser degree, but Meta’s unusually high demand for Google’s models made the squeeze hit it hardest. Due to the limits, Meta has pushed staff to use AI tokens, the units that measure AI usage, more efficiently. Google Cloud CEO Sundar Pichai has said capacity constraints held back the unit’s growth and helped nearly double its backlog quarter over quarter, even as revenue reached $20B in the first quarter.
