Meta Platforms, the company led by Mark Zuckerberg that owns Facebook, Instagram, and WhatsApp, is considering opening a division dedicated to cloud computing and competing with other US tech companies already active in this sector, such as Amazon (with Amazon Web Services), Microsoft (with Azure), and Alphabet (with Google Cloud).
META’S PLANS FOR CLOUD COMPUTING
As revealed by Bloomberg, Meta might open a cloud computing division to rent out its excess computing capacity to third parties, as well as sell access to some artificial intelligence models – Muse Spark, for example – currently hosted on its own infrastructure: this is a business model similar, in the first part, to that of so-called neocloud companies like CoreWeave.
Indeed, after Bloomberg disclosed Meta’s internal plans, CoreWeave’s shares plummeted by 14 percent, while the stock of Facebook’s parent company gained more than 9 percent, recording the biggest intraday gain since last April.
META’S “SUPER-INTELLIGENCE”
Behind Meta’s entry into the cloud computing sector is the need to secure additional revenues to support investments in artificial intelligence, which this year are expected to reach $145 billion. Zuckerberg, moreover, aims to achieve the so-called “personal super-intelligence,” i.e., the integration of AI-based assistants within the company’s platforms (Facebook, Instagram, and WhatsApp) and wearable devices (such as “smart glasses,” or smart glasses, but not only).
Amazon’s, Microsoft’s, and Alphabet’s cloud computing divisions are extremely profitable, with quarterly revenues in the tens of billions of dollars. But it is a complicated business, requiring not only data centers and high-performance microchips but also adequate software platforms, dedicated sales personnel, and efficient customer support services.
ZUCKERBERG’S WORDS
Mark Zuckerberg described the opening of a cloud computing division as “an option on the table” last May during a meeting with Meta shareholders. He revealed that “almost every week there are companies approaching us asking to start an API service [i.e., renting AI models, editor’s note], or if we have computing capacity they could buy from us at a price slightly higher than what we paid for it. We haven’t done it yet,” he explained, “because we believe we can use that computing capacity. But obviously, if we reach a point where we think we have excess, then that would be an option available to us.”
