25.06.2026
2 min
27
Google has limited Meta’s access to its Gemini AI models after being unable to provide the computing capacity the company requested. As a result, some of Meta’s AI projects have been delayed, according to the Financial Times.
According to the report, Google informed Meta that it would not be able to provide the full Gemini computing capacity the company had planned to purchase. As a result, some AI projects were delayed, and Meta employees were instructed to use AI tokens more efficiently to optimize model usage.
Meta was not the only company affected by the shortage. The Financial Times reports that several other Google customers also faced similar capacity constraints. However, Meta was hit the hardest because of its exceptionally high demand for Gemini AI models.
Despite these limitations, Meta continues to expand its AI ambitions. The company recently announced the launch of Superintelligence Labs as part of a major restructuring of its AI division. It has also pledged to invest hundreds of billions of dollars in AI infrastructure and research over the coming years.
“As the pace of AI progress accelerates, the development of superintelligence is becoming increasingly visible. I believe this will mark the beginning of a new era for humanity, and I am fully committed to doing everything necessary to ensure Meta leads the way,” Meta CEO Mark Zuckerberg said.
The shortage comes amid surging demand for AI computing resources. During Alphabet’s first-quarter earnings call, Google CEO Sundar Pichai said that limited computing capacity remains the company’s biggest constraint.
According to Pichai, despite Google Cloud revenue rising 63% to $20 billion, the business could have grown even faster. At the same time, Google Cloud’s order backlog nearly doubled from the previous quarter to $460 billion, highlighting the rapidly growing demand for AI infrastructure.
