Google caps Meta's Gemini AI use amid compute crunch

Google limited Meta's access to its Gemini AI models after Meta sought more computing capacity than Google could provide, the FT reported.ft+1
The shortfall, which began around March, disrupted Meta's AI projects including content moderation and customer support efforts.tbsnews+1
Meta has urged staff to use AI tokens more efficiently and is accelerating development of proprietary models to reduce outside dependence.firstpost

Alphabet has imposed limits on Meta's use of its Gemini artificial intelligence models after Meta sought more computing capacity than Google Cloud could provide, the Financial Times reported on Sunday.ft+1

Capacity Crunch Disrupts Meta's AI Projects

Google informed Meta around March 2026 that it could not meet the full Gemini capacity the social media company had sought to purchase, according to people familiar with the matter cited by the Financial Times. The shortfall disrupted and delayed some of Meta's internal AI projects, which relied on Google's large language models for tasks including content moderation and customer support.reuters+4

The restrictions highlight the growing strain on computing resources across the technology industry as demand for advanced AI models surges. Google CEO Sundar Pichai warned earlier this year that 2026 would be "intense," citing compute shortages as a bottleneck for products like Gemini 3.facebook

Meta Adjusts Strategy

In response to the constraints, Meta has encouraged staff to be more efficient with AI tokens — the units that measure AI usage — according to the Financial Times report. The company has also accelerated its push toward developing proprietary AI models to reduce its dependence on external providers.ap7am+2

Meta had been using Google's Gemini large language model for much of its moderation and customer support work, making the usage caps a direct operational challenge for one of the world's largest social media companies.ft

Broader Industry Implications

The episode underscores a tension at the heart of the AI boom: even the largest cloud providers are struggling to keep pace with enterprise demand for computing power. Google itself has been navigating its own capacity pressures, having recently shifted consumer Gemini subscriptions to a compute-based usage system at Google I/O 2026 to better manage resource allocation. The move from fixed prompt limits to rolling compute-based caps drew complaints from subscribers who found themselves hitting limits too quickly.androidcentral+2

Neither Google nor Meta has publicly commented on the reported restrictions.

Sources (15)