Google limits Meta's access to Gemini AI models amid high demand: report

Google has reportedly placed limits on Meta's use of its powerful Gemini artificial intelligence models, according to a Financial Times report. The restrictions emerged around March when Google, owned by Alphabet, informed Meta that it could not fulfill the full extent of the computing capacity the social media giant sought to purchase.

This shortfall has significantly impacted Meta's internal AI development, causing disruptions and delays to several projects. While other Google clients have also experienced lesser effects, Meta's exceptionally high demand for the AI models made it particularly vulnerable to these capacity constraints. The situation underscores a wider industry struggle to acquire adequate computing power to meet the escalating demand for AI services, despite substantial investments in chips and data centers.

In response to these limitations, Meta has reportedly urged its staff to use AI tokens, the units measuring AI usage, more efficiently. This measure reflects the ongoing challenge companies face in balancing ambitious AI development goals with the available computational resources. Even Google Cloud, which saw its revenue grow to $20 billion in the first quarter, acknowledged that computing power constraints hindered even greater growth and contributed to a nearly doubled backlog.

Reuters has not independently verified the report, which cited sources familiar with the matter. Both Google and Meta did not immediately respond to requests for comment outside of business hours. The report highlights the intense competition and logistical hurdles in the rapidly advancing field of artificial intelligence, where access to cutting-edge technology and sufficient infrastructure is critical for innovation and market leadership.