Google tells Meta it is going to limit Gemini access, not enough capacity available to serve AI

Google has reportedly told Meta it cannot provide all the Gemini AI capacity it wants. The restrictions are now forcing the Facebook parent to rethink how it uses AI and has even limited its employees' AI usage.

In Short

• Google reportedly limits Meta's Gemini AI access
• The restrictions come due to a shortage of computing capacity
• As a result, some of the Meta’s internal projects were delayed

Google has reportedly limited Meta's access to its Gemini artificial intelligence models after the Facebook parent sought more computing capacity than the search giant could provide. According to a FT report, Google informed Meta around March that it could not fulfil all of the Gemini AI capacity the company wanted to purchase because of infrastructure constraints, delaying some of Meta's internal AI projects.

The restrictions, which reportedly remain in place, come at a time when AI companies are facing a growing infrastructure challenge. Even companies spending tens of billions of dollars on chips, data centres and power are struggling to secure enough computing capacity to keep up with surging demand for AI models and services.

According to the report, citing people familiar with the matter, Meta has been among Google's largest Gemini customers and was hit harder than most because of its exceptionally high demand for AI computing resources. While a handful of Google's enterprise customers have also experienced similar restrictions, Meta's scale of usage has made it particularly vulnerable to the capacity crunch.

The report says Google told Meta that it simply did not have enough computing resources available to provide all the Gemini capacity the company requested. The shortage has reportedly disrupted and delayed several of Meta's internal AI initiatives.

Meta tightens employees' AI usage

As a result, Meta has also begun tightening how employees use AI internally. The company has reportedly encouraged staff to be more efficient with AI tokens—the units used to measure AI model usage—as part of a broader effort to reduce AI costs while operating under the new capacity limits.

Why is Meta using Gemini over Llama?

Gemini has reportedly become an important tool inside Meta. The company uses Google's AI models to automate safety operations, including detecting scams and removing harmful content. Gemini also powers some internal customer service tools, advertising assistants, coding workflows and productivity tasks, alongside models such as Anthropic's Claude.

According to the report, Meta initially adopted Gemini because it outperformed the company's own Llama open-source models across several enterprise use cases. However, the company has recently begun shifting some workloads to its newer Muse Spark model, which insiders reportedly believe is now competitive enough to reduce Meta's dependence on external AI providers.

Meanwhile, Google is also facing a capacity crunch

The restrictions also expose the growing infrastructure bottleneck emerging across the AI industry. While companies have largely focused on developing increasingly capable AI models, running those models at scale has become an equally significant challenge. Demand for inference computing, the processing power required every time an AI model answers a query or performs a task, has grown rapidly as businesses deploy chatbots, coding assistants and AI agents across their operations.

To meet this demand, Google has been racing to expand its infrastructure. The company reportedly signed a deal worth about $920 million a month to lease additional computing capacity from Elon Musk's SpaceX. AI startup Anthropic has also reportedly entered into a similar infrastructure arrangement.

Google itself has previously acknowledged that computing capacity remains a constraint. During the company's first-quarter earnings call in April, CEO Sundar Pichai said Google Cloud revenue crossed $20 billion for the first time but added that revenue could have been even higher if Google had enough infrastructure to meet customer demand.

"Obviously, we are compute-constrained in the near term," Pichai said at the time. "Our Cloud revenue would have been higher if we were able to meet the demand."

- Ends