Meta Platforms on Wednesday gave investors what they needed to gain renewed confidence in the company’s mounting artificial intelligence spending — and its stock. Meta is preparing to launch a cloud infrastructure business that would sell excess AI computing power and AI models to outside customers, Jim Cramer confirmed Wednesday, throwing its ring in the competitive cloud market with hyperscaler titans like Amazon Web Services, Alphabet’s Google Cloud and Microsoft’s Azure. Bloomberg News first reported on Meta’s plans. Shares of the Facebook and Instagram parent soared more than 9% Wednesday to $617 per share, among the biggest gainers in the S & P 500 . The upbeat reaction is hardly a surprise to us — Jim has recently stepped up his calls for Meta to start a cloud business, predicting the struggling stock would soar in response. “Until today, our feeling was, what the heck is Meta doing?” Jim said Wednesday on CNBC. Now, he added, “they’re going to use that [compute] power to offer a profitable enterprise to their customers.” Meta shares entered Wednesday down almost 7% for the year, trailing both the S & P 500 and the tech-heavy Nasdaq Composite, which were up 9.55% and 12.4%, respectively. It was also the second-worst performer in the “Magnificent Seven,” ahead of only Microsoft, which has been caught up in the broader “AI is eating enterprise software” narrative. Shares of Amazon and Google, the two other cloud giants, were up 5% and 14.6%, respectively. Meta has faced growing questions on its plan to monetize its enormous levels of capital spending on servers, data centers and network infrastructure. In the past, Meta has defended its investments in AI computing by saying it’s improving its advertising business for Facebook and Instagram. But it’s started to severely crimp Meta’s free cash flow, reaching levels that made some investors uncomfortable considering its narrow and economically sensitive source of revenue. In 2024, Meta’s capex totaled $37.2 billion, before rising to $69.6 billion last year. That’s projected to almost double this year, to $135 billion at the midpoint of its guidance range. For comparison, Microsoft said in April that it plans to spend roughly $190 billion on capex this calendar year. While that above Meta’s outlook, the key difference is Microsoft has a cloud business to serve. A similar defense applies to Google’s $180 billion to $190 billion in projected 2026 capex, as well as Amazon’s guidance for $200 billion . Building a cloud business, however, offers Meta another path to making money from all its AI spending, Jim explained Wednesday, which should help alleviate some of the market’s concerns and improve attitudes toward the stock even before revenue starts landing in its coffers. In other words, Meta no longer would be a one-trick pony. Better yet, this new endeavor of cloud computing has proven to be an immensely profitable business for other companies. This isn’t a complete surprise. In late May, Meta CEO Mark Zuckerberg said launching a cloud computing business was “definitely on the table.” With the stock still languishing in recent weeks, Jim argued that Meta needed to start moving in that direction. In this week’s Sunday column, he wrote : It is building out [computing] power, but for whom? We don’t know. Maybe just its own advertising model? That’s an ugh, and that’s why its stock is going down. A coherent statement from Zuckerberg right now, one that says, “We aren’t going to spend this data center money and wreck our balance sheet,” or, even better, “We are going to monetize the power by building a web services system,” — either one would get the stock out of the doldrums, which makes it a tantalizing investment still. To be sure, questions remain about Meta’s plan to sell access to computing power. For Meta to successfully compete in cloud computing, it will require far more than just owning AI data centers, according to tech analysts the Investing Club interviewed before Bloomberg’s report Wednesday morning. Tech industry analyst Ben Bajarin said investors should distinguish between two very different kinds of compute businesses. One is renting out AI infrastructure, which he called “bare metal” computing. In this case, customers would bring their own software and run it on Meta’s hardware. The other is building a full-service cloud platform complete with software, developer tools and enterprise services like AWS, Microsoft Azure and Google Cloud. “The question really is, are they offering infrastructure to third parties, or are they trying to layer software on top of that,” said Bajarin, chief executive and principal analyst at Creative Strategies, a Silicon Valley-based research firm focused on the technology industry. He also co-hosts “The Circuit” podcast, which covers semiconductors and the AI compute industry. Bloomberg’s report suggests Meta is evaluating both approaches. One proposal would resemble AWS Bedrock, by allowing developers to access AI models hosted on Meta’s infrastructure, while another would involve selling raw computing capacity similar to neocloud providers like CoreWeave or, more recently, SpaceX. Elon Musk’s rocket-and-AI company last month struck a deal with Google in which Google will pay $920 million a month for additional computing power. SpaceX, which has built a massive data center near Memphis, Tennessee, also a similar deal with Anthropic. Bajarin said the timeline of when Meta’s cloud business materializes depends on how ambitious its cloud plans are. If the company rents out excess AI infrastructure, the offering could arrive quicker because customers would supply their own software. On the other hand, building a full-fledged cloud platform like AWS or Google Cloud would take much longer. That’s because he said building a cloud business to serve outside customers is harder than building data centers for internal workloads. Bajarin said it requires software that allows customers to deploy workloads on its infrastructure, an area in which established cloud providers have spent years investing. Paul Meeks, head of technology research at Freedom Capital Markets, said Meta’s investment-grade balance sheet gives it a major edge over newer AI infrastructure providers that have relied heavily on debt to fund expansion . What Meeks questioned, however, is whether AI companies would want to host sensitive workloads on infrastructure owned by a competitor that’s also building its own AI models and applications. He outlined the way AI labs like OpenAI and Anthropic may be thinking about it: “If Meta has a product that is competing with us, then people will hesitate to buy their cloud services,” said Meeks, whose covered tech for decades. At the same time, Bajarin argued that demand for AI compute remains so strong that customers “will take compute wherever they can get it,” which would work if Meta took the bare-metal approach. And if it pursues a full-fledged cloud service, Meta has relationships with tons of businesses that use Instagram, Facebook and WhatsApp — something we believe sets them up to be potential cloud customers. Bottom line We’re pleased that Meta is taking steps to explore ways to commercialize its AI infrastructure and show Wall Street that it’s sensitive to investor concerns. At the same time, bigger questions remain about how ambitious the company intends to be — whether it wants to be just another supplier in a compute hungry market or if it builds a full-service cloud platform. Either way, the move is another welcome step in Meta’s efforts to turn its enormous AI investments into meaningful long-term returns for investors. (Jim Cramer’s Charitable Trust is long META, AMZN, GOOGL. See here for a full list of the stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
