Google Reportedly Limits Meta's Access to Gemini AI Models Amid Surging Demand

28 June 2026

The race to dominate artificial intelligence has become so intense that even the world's biggest technology companies are beginning to face resource shortages. According to a recent report, Google has placed limits on Meta's use of its Gemini artificial intelligence models after the social media giant requested more computing capacity than Google was able to provide. The development highlights the growing strain on AI infrastructure as companies compete to build increasingly powerful systems while demand continues to outpace available resources.

The reported restrictions were introduced around March after Meta sought significantly greater access to Google's Gemini models for a variety of internal projects. While Google has continued supplying AI services to Meta, the company was reportedly unable to satisfy the full level of computing power requested. As a result, several of Meta's AI initiatives experienced delays while engineers adjusted to the reduced capacity.

The situation illustrates one of the biggest challenges currently facing the artificial intelligence industry. Although companies are investing hundreds of billions of dollars into new data centers, advanced processors, and cloud infrastructure, demand for AI computing power continues growing even faster. Training and operating large language models require enormous quantities of specialized hardware, making access to computing resources one of the industry's most valuable assets.

According to the report, Meta had been using Google's Gemini models for several important internal functions, including software development, customer service automation, and safety related systems. With access becoming more limited, the company reportedly instructed employees to use AI processing tokens more efficiently in order to maximize the available computing resources. Those internal measures were intended to reduce unnecessary AI usage while keeping critical projects moving forward.

The reported capacity shortage affected more than just Meta. Other Google Cloud customers also experienced limitations, although the Financial Times reported that Meta was impacted more heavily because of the unusually large amount of computing power it had requested. The situation underscores how even the largest cloud providers are struggling to satisfy unprecedented levels of demand generated by the ongoing AI boom.

Google has acknowledged that strong customer demand has placed pressure on its cloud infrastructure. Earlier this year, Alphabet Chief Executive Sundar Pichai said the company's cloud division had achieved more than $20 billion in quarterly revenue while simultaneously experiencing capacity constraints that prevented even greater growth. He also noted that Google Cloud's order backlog had expanded significantly, reflecting demand that currently exceeds available computing resources.

For Meta, the reported restrictions arrive during one of the company's most ambitious periods of AI investment. Chief Executive Mark Zuckerberg has committed to spending approximately $600 billion on infrastructure through 2028 as Meta works to expand its artificial intelligence capabilities across products including Facebook, Instagram, WhatsApp, and its growing family of AI assistants. The company is also developing increasingly advanced proprietary models as it seeks to reduce dependence on outside technology providers.

Industry analysts believe the reported limits may encourage Meta to accelerate development of its own internal systems. According to the Financial Times, the company has already begun shifting more projects toward its in house Muse Spark AI model, reducing reliance on external providers such as Google whenever possible. That strategy reflects a broader trend among major technology firms, many of which are simultaneously partnering with competitors while building rival technologies of their own.

The report also highlights the unusual dynamics shaping today's artificial intelligence industry. Companies that fiercely compete in consumer products and AI development often rely on one another behind the scenes for cloud infrastructure, specialized chips, and advanced models. These partnerships allow innovation to move more quickly but also create situations where competitors must share limited resources during periods of extraordinary demand.

Neither Google nor Meta publicly commented on the report at the time it emerged. Reuters said it was unable to independently verify every aspect of the Financial Times report, though the publication cited people familiar with the matter.

The reported restrictions serve as another reminder that the future of artificial intelligence depends not only on breakthrough algorithms but also on the enormous physical infrastructure required to support them. Data centers, graphics processors, networking equipment, and electricity have become just as important as software in determining which companies can lead the next phase of AI development.

As competition intensifies, technology companies are expected to continue investing aggressively in expanding computing capacity. Until that infrastructure catches up with demand, however, even the world's largest AI developers may find themselves competing for one of the industry's most valuable commodities: access to enough computing power to fuel the next generation of artificial intelligence.