Google has reportedly limited the amount of Gemini AI computing capacity available to Meta after surging demand for artificial intelligence services placed growing pressure on the company's cloud infrastructure.

According to reports, Google informed Meta in March 2026 that it would be unable to provide the full level of Gemini AI resources requested. The decision reflects the increasing strain on global AI infrastructure as technology companies race to build and deploy more advanced generative AI models.

The reported restrictions come after demand for Google's AI services accelerated rapidly over the past year. API requests for Gemini reportedly more than doubled between March and August 2025, placing significant pressure on available computing resources and data center capacity.

Meta, which had planned to expand its use of Google's AI infrastructure, was forced to revise its strategy after learning that the requested capacity would not be immediately available. The company reportedly began optimizing its internal AI workloads while also exploring alternative infrastructure providers to support future growth.

As part of that effort, Meta later reached a multibillion-dollar agreement with Nebius Group to secure additional AI computing infrastructure beginning in 2027. The deal is expected to provide the company with expanded access to high-performance computing resources needed to train and operate large AI models.

The development highlights one of the biggest challenges facing the artificial intelligence industry today. While companies continue introducing increasingly powerful AI systems, the availability of graphics processing units (GPUs), cloud servers and data center infrastructure has struggled to keep pace with demand.

Google has previously acknowledged that infrastructure capacity remains one of its biggest operational challenges. During recent financial results, the company reported continued growth in Google Cloud revenue while noting that demand for AI computing services continues to exceed available capacity in several areas.

The rapid adoption of generative AI has intensified competition among major technology companies for access to advanced computing resources. Cloud providers are investing billions of dollars in new data centers, specialized AI chips and networking infrastructure to support growing enterprise demand.

The reported capacity constraints also demonstrate that infrastructure availability has become as strategically important as AI model development itself. Companies building frontier AI systems increasingly depend on reliable access to large-scale computing resources to maintain product development and serve expanding customer bases.

Despite the reported limitations, both Google and Meta continue to invest heavily in artificial intelligence. Google is expanding its cloud infrastructure and AI services, while Meta is increasing spending on data centers, custom AI hardware and large language model development.

The situation underscores the growing importance of computing power in the global AI race. As organizations deploy larger models and millions of users adopt AI-powered services, securing sufficient infrastructure has become a key competitive advantage for technology companies.

With demand continuing to outpace available resources, industry observers expect cloud providers and AI developers to accelerate investments in next-generation infrastructure over the coming years, making computing capacity one of the defining factors shaping the future of artificial intelligence.