Meta is seeking to establish its dominance in artificial intelligence and has signed an agreement to use Google Cloud’s cloud computing services worth at least $10 billion over six years, according to a Bloomberg reports.
The deal would expand Meta’s capabilities as the company continues to promote artificial intelligence tools and services across its platforms, including Facebook, Instagram and WhatsApp. AI requires massive amounts of computing resources and increased data center bandwidth. These data centers require more and more electricity and water to power and cool the internal hardware.
Meta is also spending large sums of money building its own AI data centers, including a 4 million square foot facility in Louisiana called hyperionopening planned for 2030.
Meta and Google declined to comment on the details of the deal or confirm that an agreement had been reached.
Meta’s major advances in artificial intelligence
In an effort to hire a superstar team to work on its AI projects, including the Llama AI platform, Meta has raised hundreds of millions of dollars. However, recent reports suggest that the recruitment is due to artificial intelligence. can be paused when the company reevaluates its strategy.
Additionally, Meta continues to introduce new AI tools and features, including AI translations for Facebook and Instagram.
Meanwhile, a deal with Meta could cement Google Cloud’s status as what Bloomberg calls a “one-stop shop” for AI services. In the month of June, OpenAI works together with Google Cloud use their data centers.
Meta has its own data centers and has even invested in nuclear power to meet its data goals, but its needs seem to be growing faster than they can keep up.
“Across the industry, we’re seeing AI workloads drive higher performance and compute capacity demands that exceed even the most demanding data center construction schedules,” said Chuck Marvin, Chief Commercial Officer at Thunderhead Power Solutionswhich has a close collaboration with data centers.
While AI workloads are one of the reasons the data center industry is booming, Marvin says it’s not the whole story.
“Yes, AI workloads require a lot of power, especially training large language models and running large-scale inference, but this is happening at the same time as significant growth in enterprise cloud adoption, use of edge computing and digital transformation initiatives that are already straining data center capacity,” says Marvin.
Artificial intelligence increasingly uses clusters of power-hungry GPUs, but other cloud services, such as video streaming, are also spreading around the world. These demands are converging and pushing supply networks to their limits.
“The result is that data center operators are struggling to secure power and capacity for all their use cases, not just AI,” he said.
What is the future of Google Cloud?
While AI is not the only driver of data center growth, it is still important. Google Cloud is uniquely positioned to provide these types of services, Marvin said.
Notably, the company has developed its own TPU chips and has extensive AI experience handling large-scale search and YouTube applications. It also built its infrastructure specifically for the massive data movements that AI requires, Marvin says. “Google Cloud has gained traction by focusing on its core competencies in artificial intelligence and machine learning… and giving enterprise customers access to the same advanced infrastructure that powers Google’s services,” he said.
There is plenty of demand for multiple winners, including competitors like Microsoft and Amazon, but Google has done a good job of building an AI-powered infrastructure. “Especially for AI-intensive workloads, Google offers technological advantages that are hard to compete with,” Marvin said.
