United States limits Meta access to Google Gemini
Tap title to read story
Google has significantly restricted Meta Platforms’ access to its Gemini artificial intelligence models due to a severe shortage of computing capacity, according to reporting first highlighted by the Financial Times. Around March 2026, Google told Meta it could not supply the full Gemini capacity Meta had requested, with the social media company particularly affected because of its unusually high demand. Meta had been using Gemini extensively for safety and content moderation tasks, where it reportedly outperformed Meta’s Llama models. In response, Meta is tightening internal usage of AI tokens and accelerating a shift to its in-house Muse Spark model for key moderation workloads.
More context
Timeline
- Around March 2026, Google notifies Meta
- March 2026, Gemini capacity request partially denied
- Spring 2026, Meta projects disrupted by limits
- Spring 2026, Google rations compute among clients
- Spring 2026, Meta tightens internal AI token usage
- Spring 2026, Meta shifts workloads to Muse Spark
- Spring 2026, Llama underperforms Gemini on moderation
- Spring 2026, infrastructure shortages constrain AI inference
Why it matters
If you use Meta's platforms, this could impact your experience. The shift to Muse Spark might affect content moderation, possibly letting more harmful content slip through. Keep an eye on your feed and report any inappropriate content.
JQJO Editorial Desk · Human reviewed · AI assisted. Report a correction.
