Google Limits Meta Access to Gemini AI Due to Compute Shortages

Google has imposed strict limits on Meta’s access to its Gemini AI models, citing extreme demand that has outpaced available compute capacity. This decision, reported on June 28, has resulted in significant disruptions to Meta's internal artificial intelligence projects. While Google Cloud achieved a robust $20 billion in revenue for the first quarter ending in March, CEO Sundar Pichai acknowledged that severe computing power constraints hindered even higher growth, with the backlog doubling quarter-on-quarter.

The challenge is not unique to Meta; numerous companies relying on Google’s infrastructure are facing similar restrictions as big tech firms scramble for limited hardware resources. The industry is currently defined by massive capital expenditure on AI-specific chips and data center expansion. Beyond Google, other firms are taking defensive measures. For instance, Microsoft has instructed its internal units to wind down usage of Anthropic’s Claude Code in favor of its own proprietary CLI tools, reflecting a wider trend of optimizing resource efficiency.

The recent development highlights a critical bottleneck in the AI ecosystem: the scarcity of high-performance computing power. Companies are increasingly adopting 'tokenmaxxing' strategies, encouraging staff to be hyper-efficient with their AI resource usage to mitigate costs and manage hardware scarcity. Meta, for example, is now tying employee performance to AI tool usage while simultaneously requiring more efficiency per token. As global demand for AI services continues to climb, the ability to secure and manage compute infrastructure will likely remain the primary factor determining growth trajectories for major tech players in India and abroad.