The Financial Times reported on June 28, citing three sources familiar with the matter, that Google informed Meta around March that it could not provide the full computing capacity Meta sought to purchase. This decision came after Meta requested more computing power than Google could supply.
The restrictions are still in place, causing delays in some of Meta's internal AI projects. Additionally, as part of a cost-cutting initiative, Meta is encouraging employees to use AI tokens more efficiently. AI tokens are units that measure the usage of AI models.
Google's usage limits have affected not only Meta but also other clients. However, Meta has been particularly hard hit due to its high demand for Google's models. The Financial Times noted that Google's decision to limit access for major clients highlights the growing infrastructure pressures and bottlenecks within the AI industry.
Global tech giants are investing billions of dollars in securing semiconductors, data centers, and power, yet they are struggling to obtain enough computing power to meet the demand for high-performance AI models and services.
Google is also accelerating efforts to secure additional computing capacity. According to the Financial Times, the company signed a contract earlier this month with Elon Musk's SpaceX to lease computing capacity for $920 million per month.
Sundar Pichai, Google's CEO, announced during the first-quarter earnings call in April that Google Cloud revenue surpassed $20 billion for the first time. He also mentioned that while contracts have been signed, the unfulfilled cloud contract backlog has nearly doubled from the previous quarter, exceeding $460 billion.
Pichai stated, "We are facing computing constraints in the short term, and if we could meet demand, our cloud revenue would have been even higher."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.