The Financial Times reported that Google has imposed limits on Meta's use of Gemini AI models.
Google Limits Meta's Use of Gemini Due to Compute Shortages, FT Reports
This article is a translation. Read the Japanese original
According to the publication, Google informed Meta around March that it could not provide all the requested computing power. This shortage reportedly caused disruption and delays for some of Meta's internal AI projects.
While other customers were also affected, the impact was not as severe as it was for Meta. Meta is said to have been particularly hard hit due to its extremely high demand for Gemini.
Reuters could not immediately confirm this report. According to sources, Google and Meta did not respond to requests for comment as it was outside of business hours.
In response to the limits, Meta urged its employees to improve the efficiency of their AI token usage. Tokens are the units used to measure AI consumption.
Although companies are investing massive sums into chips and data centers, they are struggling to secure the computing power necessary to meet the growing demand for AI services.
Google Cloud's revenue grew to $20 billion in the quarter ending in March. However, CEO Sundar Pichai stated that further growth has been hindered by compute constraints, and the backlog has nearly doubled compared to the previous quarter.
Source: Google limits Meta's use of its Gemini AI models（HN 162pt・72コメント） (HN Search (backfill), 2026-06-28)