The best defense is a good attack. This old sports adage seems to describe what Meta Platforms CEO Mark Zuckerberg hopes to accomplish by launching his cloud business. Shares in the parent company of Facebook and Instagram soared nearly 9% last Wednesday, the day the company confirmed to Jim Cramer that a long-talked-about cloud service was in the works. According to Bloomberg, which first reported the news, Meta is debating whether to offer access to AI models hosted on its infrastructure or sell access to its raw computing power. A week later, few details have been released about what exactly Meta plans to do, and the stock is down several dollars per share from its closing price on the day the news broke. The bull case for the meta cloud business is that by going on the offensive and creating new revenue streams, it can also provide the necessary defenses to prevent overbuilding of AI infrastructure. Concerns about these spending became evident on the night of April 29, when Meta’s first quarter 2026 fiscal year backlog was hit by a heavy sell-off. The Street newspaper evaluated Meta’s capital investment intentions more harshly than its hyperscalers (Amazon, Microsoft, Alphabet). Monetizing that spending depends almost entirely on internal demand for computing. The big three public clouds use only the compute you need and rent out excess capacity. In the bearish case, the question is whether the metacloud is exhibiting too much compute because the enterprise has exceeded the required amount. Let’s explore the arguments and conclusions for each. The bearish case The bearish case will focus on whether Meta can generate a sufficient return on invested capital (ROIC) for the massive AI spending it has already undertaken and its $125 billion to $145 billion capex plan for fiscal 2026. That’s an increase in capital spending from the previous range of $115 billion to $135 billion, and even at the low end is higher than the expected $122.64 billion, according to FactSet. Laura Martin, a longtime Needham analyst, said in a note Monday that Meta is moving into the cloud business because its AI infrastructure is so overbuilt that it won’t need all the compute it will generate in its 2026 capital spending plan. Martin and his team also said that Meta would find it difficult to enter the cloud business “now with deep-pocketed competitors such as AWS, Google Cloud, and MSFT Azure.” They also expect a good return on invested capital (ROIC) as Meta “shifts its focus from its 70% margin core advertising business to its 35% margin cloud business.” Meta is flooding the zone with all sorts of AI announcements, including Tuesday’s release of Muse Image, a new AI model for creating images aimed at attracting creators and advertisers. We noted last month that all of these efforts, from low-cost smart glasses to enterprise tools for businesses to plans to build a market-predicting app and a key partnership with Qualcomm, have done little to help Meta stock out of its turmoil. META Year-to-date Mountain Meta Platforms YTD Meta has the second-worst year-to-date stock price performance among its hyperscaler peers, down more than 8%. Microsoft, which has had its own problems over concerns about software interruptions, is down about 20%, while Amazon is up nearly 5%. Alphabet has been a big winner so far, up more than 15% since the beginning of the year. Meta also has the lowest price multiple, at 17.7 times expected earnings over the next 12 months, compared to Amazon’s 25.5 times, Microsoft’s 19.6 times, and Alphabet’s 24.9 times. Analysts at Canaccord Genuity believe the Metabare incident has gone “too far.” They said in a note on Monday that “accelerating advertising business, emerging subscription tiers, and external markets that are now paying visible fees for capacity make it increasingly difficult to justify META’s discount to the rest of Mag7.” In the case of a bull Let’s talk about the case of a bull. JPMorgan estimated that every gigawatt of metacomputing power provided by cloud businesses could generate $20 billion in annual revenue and several dollars in earnings per share (EPS). Meta’s fiscal 2025 revenue rose 22% to nearly $201 billion. FY2025 EPS decreased by 1.6% to $23.49. (Compute is measured in gigawatts because power is the limiting factor in data centers. So 1 GW of compute means an AI infrastructure that consumes 1 GW of continuous energy, or enough energy to power up to 1 million homes.) Meta made us realize that right now, we need all the compute we can get. While cloud businesses can create more internal capacity constraints, this move provides a necessary hedge for companies to streamline spending. We know that the CEOs of these technology companies believe that the risk of underspending outweighs the risk of overspending, and as a result, they are willing to take on the risk of overspending. Meta’s cloud plans provide an off-ramp should the day come when Meta uses more than a few gigawatts of excess capacity. Additionally, Meta can significantly expand its computing demands by exposing its infrastructure to the world. That demand will also become more diverse, as it will come from many different companies in many industries, each in different phases of their business cycles and with their own AI demand profiles. This is why public cloud business models are so successful. The club’s view As meta-investors in a club’s portfolio, we need to consider both sides of the argument. But perhaps it doesn’t have to be one or the other. Perhaps Meta is still internally compute-constrained. It may also be true that Mr. Zuckerberg has decided that if he wants Wall Street to support him and his stocks, he must provide a solid rationale for all his spending. Public cloud businesses achieve both goals at the same time. Why is that an attractive proposition? Consider what we just learned about the relationship between Meta and Alphabet. Google’s parent company told Meta that it needed to limit its use of Gemini because it did not have the capacity to meet Meta’s demand. Meta is at its mercy by being a renter rather than a landlord. If Alphabet internally determines it needs more, it could limit supply to customers. Wall Street doesn’t seem to look favorably on companies that have the option of renting their computing, instead hoarding it for internal efforts — especially when they aren’t confident that their internal plans will generate as much or as much short-term profit as simply renting. That’s part of the reason Microsoft has been underperforming this year. (Microsoft also has too much exposure to enterprise software, but as a group it has been hammered by concerns about AI disruption.) In Meta’s case, the company is likely to end up reallocating and renting out in-house computing that is deemed less profitable. In other words, it would do the opposite of what some investors see as a problem with Microsoft. That should please the streets. We see this as a strong move. Because it performs offensively while being incredibly effective defensively for the dizzying levels of capital investment required to ensure no one gets left behind. (Jim Cramer’s charitable trusts are long META, AMZN, MSFT, GOOGL. See here for a complete list of stocks.) As a subscriber to Jim Cramer’s CNBC Investment Club, you will receive trade alerts before Jim makes a trade. After Jim sends a trade alert, he waits 45 minutes before buying or selling stocks in his charitable trust’s portfolio. If Jim talks about a stock on CNBC TV, he will issue a trade alert and then wait 72 hours before executing the trade. The above investment club information is subject to our Terms of Use and Privacy Policy, along with our disclaimer. No fiduciary duties or obligations exist or arise from your receipt of information provided in connection with the Investment Club. No specific results or benefits are guaranteed.
