The $10 billion, six-year cloud deal between Google and Meta shows the social media giant plans to continue scaling all cloud providers as it works towards its super intelligence goals.
Last week, several media sources revealed that Meta had made deals with Google primarily to use AI infrastructure.
The news comes after Meta revealed it in its latest revenue report. Revenue for the two quarters was $475.2 billion, up 22% from the previous year. Costs and expenses for the second quarter were $27.07 billion, an increase of 12% year-on-year. The company expects total costs for 2025 to be between $11.4 and $118 billion. The social media giant said higher infrastructure and pay costs were primarily driving second quarter expenses.
“Our main focus is continuing to invest capital in our business, and infrastructure and talent are our number one priority,” Chief Financial Officer Susan Lee said over the phone. “We also expect the development of key AI infrastructures to be a central advantage in developing the best AI models and product experiences, so we look forward to a significant increase in investment in 2026 to support the work.”
Meta must be scaled
This transaction is primarily about scaling, as that is the goal.
They are looking for computational scaling, they are looking for infrastructure scaling.
Chirag dekateAnalyst Gartner
“They are looking for computing scaling. They are looking for infrastructure scaling,” said Chirag Dekate, a Gartner analyst. “What they need is raw computing power delivered in a reliable and predictable way.”
Google's contracts usually work with all cloud hyperscalers, so they also have a brand with Meta.
“To operate your own data center and all three hyperscalars, you need meta to ensure that your AI platform works equally on all of this infrastructure. This is important when it comes to ensuring the highest possible scalability.” “Towards superintelligence, the degree of scalability as much as possible is a competitive advantage for the meta over each hyperscalar.”
Google's Benefits
Dekate said this is a win for Google, despite this being a strategic move that will help the meta expand it.
“It's a solid test of its AI and native leadership,” he said. He added that Meta also chose Google for its networking stack.
“They innovated their own core infrastructure stack that allows them to scale across data centers and allow software-defined networking at scales that others can't handle,” he said.
Google and Meta are competitors when it comes to AI models and other aspects, such as advertising, search and the consumer market, but infrastructure levels are different issues, Dekate said.
“The infrastructure layer emphasizes the need and desire to continue to leverage data center capabilities and continue to leverage high efficiency, high reliability, high capabilities and AI computing resources,” he said.
It's not just AI vendors affiliated with Google. Openai added Google as a cloud provider at the beginning of the summer.
“The fact that Google's competitors use infrastructure is not only a testament to infrastructure leadership, but also security, isolation, and how Google competes in space,” Dekate said.
He added that having a good infrastructure or cloud provider is also necessary for Genai Race and Meta to drive super intelligence.
“In races where meta, Gemini and open ry are on, the direction of spread is first consumers and the enterprise continues, so it's actually heading towards the consumer market,” he said.
To be the best in the consumer market, vendors need a good infrastructure, leverage inference mechanisms and train the best AI models.
Esther Shittu is an Informa TechTarget News Writer and Podcast host that covers artificial intelligence software and systems.
