Meta’s AI Monetization Plan Sparks Rally
Meta Platforms Inc. posted its best weekly stock performance since early 2024, with shares rising roughly 15% and an additional 6% on Friday alone. The surge was not fueled by its core advertising business, which generates the vast majority of revenue, but rather by a strategic pivot: the launch of Meta Compute, a new initiative to sell artificial intelligence computing capacity and models to external customers. This move directly addresses investor anxiety over Meta’s massive capital expenditures on AI infrastructure, which had left the stock flat while the Nasdaq-100 climbed 18% over the past year.
The Catalyst: Meta Compute
Meta Compute is designed to rent out spare AI computing power—GPUs and proprietary MTIA chips—along with access to Meta’s large language models and image generation tools. Chief Executive Officer Mark Zuckerberg had previously hinted at this direction, stating that an AI cloud business "makes sense." The announcement coincided with the release of Meta’s Muse Image model and news that Meta’s own MTIA chip is entering production, reducing dependence on Nvidia. For investors, the plan transforms a colossal cost center into a potential revenue stream, providing a tangible path to returns on the tens of billions Meta has poured into AI data centers.
Key Facts Behind the Rally
- Stock performance: Meta shares rose ~15% over the week, the best week since early 2024, with options volume exceeding three times the 30-day average. 78% of $1.8 billion in options premium was tied to calls.
- Analyst estimates: Wolfe Research projects that for every gigawatt of compute monetized at a ~$25 billion annual rate, earnings per share could increase by roughly 20%. Price targets cluster in the low-to-mid $800s over twelve months, with a bearish scenario around $720 and bullish near $869.
- Customer zero? Meta Compute has not secured any external customers yet. Meta has never operated a cloud business for outside clients, placing it at a significant disadvantage against Amazon Web Services (AWS), Microsoft Azure, and Google Cloud—three hyperscalers with decades of operational experience, established sales teams, and trust.
- Job cuts persist: Even as Meta celebrates this new revenue opportunity, it has cut over 8,000 jobs while posting record revenue, a reminder of ongoing cost rationalization amid record AI spending.
Expanding the Context: Meta’s AI Spending Spree
Meta’s capital expenditures for 2024 are expected to exceed $35 billion, with a significant portion allocated to AI infrastructure—building out data centers, acquiring Nvidia H100 and B200 GPUs, and developing custom silicon like the MTIA chip. This spending spree had unsettled Wall Street, as Meta’s previous attempts to monetize AI (e.g., through advertising improvements) were often overshadowed by high costs. The launch of Meta Compute provides a clearer narrative: the same infrastructure can be repurposed for external customers, creating a new line of business akin to AWS’s origins as an internal tool turned public service.
The Hyperscaler Challenge
Meta faces immense competition. AWS, Azure, and Google Cloud together dominate the global cloud market, with combined market share exceeding 65%. They offer not only compute but also data storage, networking, security, and enterprise support—services Meta has not developed for external use. Meta’s strength lies in its research and open-source AI models (e.g., Llama), but renting out raw compute requires a different skill set: customer support, SLAs, compliance certifications, and multi-tenant isolation. Analysts remain skeptical that Meta can quickly build these capabilities while simultaneously managing its core social media business.
What the Rally Ignores
The market’s enthusiasm may be betting on a narrative rather than proven execution. Meta Compute has yet to announce any customers; the plan exists only as a press release and internal roadmap. Furthermore, selling excess capacity can be interpreted as a sign that Meta overinvested in AI hardware—a potential admission of miscalculation. Meanwhile, the job cuts underscore a tension between workforce cuts and massive infrastructure spending, raising questions about strategic coherence. Until Meta signs its first cloud contract, the rally remains speculative.
Historical Parallels
The situation mirrors Amazon’s early move to create AWS in 2006, but with critical differences: Amazon had years of internal experience running a highly efficient retail platform and could leverage its engineering culture. Meta, despite its engineering prowess, has primarily focused on consumer applications and advertising. Its foray into enterprise cloud services would require a culture shift and heavy investment in sales and support—costs not yet factored into analyst projections. Additionally, current hyperscalers are aggressively incorporating AI into their offerings, making it harder for a latecomer to gain traction.
Financial Implications
If Meta Compute succeeds, it could significantly diversify revenue beyond advertising, which currently accounts for over 98% of total sales. Wolfe’s estimate of a 20% EPS lift per gigawatt assumes high utilization and pricing comparable to existing cloud providers. However, pricing wars and capacity glut could compress margins. Meta’s own MTIA chip may offer cost advantages over Nvidia GPUs, but custom silicon development is risky. The company must also balance internal demand for AI compute (e.g., for training models and content recommendation) with external commitments.
The human cost of Meta’s AI pivot remains high. Layoffs have affected thousands of employees across recruiting, legal, and marketing departments, even as Meta hires top AI researchers. This duality highlights the company’s determination to streamline operations while betting big on a technology that may take years to pay off. The question for investors is whether the potential returns justify the ongoing social and operational disruption.