Google has restricted Meta’s use of its Gemini artificial intelligence models due to capacity constraints. This limitation, in place since around March, has led to Google being unable to fulfill Meta’s full AI computing needs, causing disruptions and delays in Meta’s internal AI projects. Meta, a major customer of Google’s AI services, faced challenges due to its high demand for Gemini models, prompting the need for more efficient AI resource usage within the company.
The strain on computing resources in the tech industry has been highlighted by Meta’s experience, with other Google customers also encountering limitations in accessing computing resources, albeit to a lesser extent than Meta. This situation reflects the overall pressure faced by tech giants as they strive to enhance their artificial intelligence capabilities amid a persistent shortage of AI computing power.
To address the capacity constraints, Meta has encouraged its employees to optimize AI resource utilization, including reducing the consumption of AI tokens used to measure and manage generative AI model usage. Despite significant investments in data centers and advanced chips, the demand for AI computing power continues to outstrip the available supply, posing challenges for companies like Google and Meta.
Acknowledging the challenges posed by capacity limitations, Google reported a rise in Google Cloud revenue to $20 billion during the first quarter. However, CEO Sundar Pichai noted that these constraints hindered even stronger growth and contributed to a notable increase in the cloud division’s backlog. The restrictions on Meta’s access to Gemini models underscore the emerging bottleneck in the tech industry due to shortages of AI infrastructure, as companies ramp up investments in generative AI and cloud computing.
