Can Meta Compute Challenge AWS and Google Cloud?

Meta is announcing plans to launch a cloud infrastructure business, a move that will see the technology company compete with established services which include Amazon Web Services, Microsoft Azure and Google Cloud. It is designed to offer external customers direct access to AI models and computing capacity.

An internal group called Meta Compute will be instrumental in these efforts and has been created to oversee the buildout and operations of its AI infrastructure.

Meta Compute is run by several company executives, including Santosh Janardhan, Meta’s Head of Infrastructure; Daniel Goss, an executive within Meta’s Superintelligence Labs AI Unit; and Meta President Dina Powell McCormick. 

One option under consideration for the computing segment is a service model which will allow outside developers to pay to run queries against AI models on infrastructure that Meta owns and operates. These models include the proprietary gen AI model belonging to Meta, which is named Muse Spark.

Additionally, the company will create a separate avenue to rent out raw GPU capacity directly to Meta customers.

Selling computing access to external developers

Speaking to shareholders in May, Mark Zuckerberg, CEO at Meta, said the idea of selling computing access is something the company has been considering. Mark added that the idea is ‘definitely on the table’.

“Almost every week there are different companies that come to us from the outside asking us to both stand up an API service or asking if we have compute that they could buy from us at some premium to what we've bought it at,” Mark said. 

He explained that no outside deals have been made because internal demand has taken up the available capacity. However, the CEO went on to say that, if overbuilding occurs, Meta will move to sell the extra capacity externally.

Investing into AI superintelligence infrastructure

Superintelligence has become a major priority for Meta as it invests hundreds of billions of dollars into AI infrastructure. Meta has also reached large capacity agreements with CoreWeave, Google and Oracle.

Some investors have questioned how operations like this will translate into revenue. Despite these doubts, the potential cloud business could prove a profitable venture for the organisation.

Compute supply constraints that have been a key point of discussion first came into focus when Google restricted access to its Gemini AI. Google says it was unable to meet the demand from Meta for AI compute.

This supply constraint delayed some internal AI efforts at Meta. The restriction resulted in Meta asking its own employees to reduce their AI token consumption.

Muse Spark has since taken on a large share of the work initially handled by Gemini. This shift comes as Meta starts to develop and grow its own in-house AI capabilities.

The company first unveiled the model in April, but it is yet to release it to developers. A confirmed release date for the AI model is yet to be announced, The Wall Street Journal reports.

Meta is projected to spend as much as US$145bn on AI infrastructure this year. This expenditure represents a significant portion of the average spend of US$700bn on the technology across the tech industry.