Meta shares fell after Q3 earnings as investors worried about surging AI spending. With 2024 capex hitting $38-40B and 2025 climbing, returns remain unclear.
Meta shares fell in after-hours trading on October 30, 2024, even as the company delivered strong third-quarter results. The reason: investors reacted to Meta’s escalating artificial intelligence spending plans. The company reported $40.6 billion in revenue and $15.7 billion in net income, but raised its 2024 capex forecast to $38–40 billion. Worse, executives warned that 2025 capital expenditures would be “significantly higher.” The tension between strong fundamentals and costly AI bets now defines Meta’s market narrative.
Meta’s Q3 2024 earnings were objectively strong. Revenue grew 19% year over year to $40.6 billion, while net income climbed 35% to $15.7 billion. The company’s advertising business continues to benefit from AI-enhanced targeting, user engagement across its apps, and steady demand from global marketers.
Yet the stock declined in extended trading. Investors shifted their attention from the quarterly scorecard to the mounting cost of AI infrastructure.
Key figures from the Q3 report:
Chief Financial Officer Susan Li delivered the critical warning during the earnings call. “We expect capital expenditures in 2025 to be significantly higher than 2024,” she said. It was those words that triggered much of the post-earnings selling.
Capital expenditures cover the physical infrastructure needed to support AI: data centers, advanced graphics processing units, network equipment, and energy systems. Meta has been investing heavily to expand its AI compute capacity, and the spending trajectory is accelerating sharply.
The market’s concern is straightforward. Today’s AI costs are certain, but the revenue they will generate is not. Meta has not provided detailed projections for AI-related sales, leaving investors to estimate when the multi-billion-dollar outlays will produce measurable profit.
The anxiety shows up in several specific questions:
There is also the precedent of the metaverse. Following Meta’s heavy investment in virtual reality and the metaverse starting in 2021, the company experienced massive operating losses in its Reality Labs division, and the payoff remains limited. Although Zuckerberg frames AI as a fundamentally different opportunity, investors are understandably wary of another long-duration bet.
Meta’s recent cost discipline adds another layer. In 2023, the company reduced headcount and streamlined operations, which lifted margins. The current AI investment cycle is now absorbing some of those gains, raising the risk that 2025 margins will come under pressure.
CEO Mark Zuckerberg was unambiguous in defending the AI spending strategy. His view: in a transformative technology cycle, the cost of moving too slowly far outweighs the cost of moving aggressively.
“I’d rather risk building capacity before it is needed rather than too late,” Zuckerberg said.
His argument is rooted in Meta’s history. The company faced similar criticism when it invested heavily in Stories and Reels, but those formats eventually drove meaningful advertising revenue. Zuckerberg expects AI to follow a similar arc—only at a far greater scale.
The counterargument is equally clear. AI infrastructure is far more expensive than adding new product features to an existing platform. Oversupply could also deflate the value of AI services across the industry. If Meta builds vast compute capacity and demand grows slowly, much of that capital could be underutilized.
A notable development in the Q3 call was Zuckerberg’s announcement that Meta would begin selling its AI tools to outside companies for the first time. This is a strategic shift from internal use toward direct monetization of Meta’s AI research and infrastructure.
The company has already released its Llama family of large language models as open-source projects. The new enterprise offerings go further, providing commercial support, managed deployment, and integration with Meta’s platforms.
What Meta now offers the business market:
This positions Meta in direct competition with enterprise AI players like OpenAI, Microsoft Azure AI, and Google Cloud Vertex AI. Meta’s open-source approach and extensive consumer ecosystem give it differentiators, but the enterprise market is competitive and demanding.
For technology professionals, the implication is practical. Meta is becoming another vendor in the AI stack. Those evaluating AI providers should assess Meta’s enterprise offerings on cost, model performance, ease of integration, and support quality.
Meta is hardly the only company spending heavily on AI infrastructure. Microsoft, Alphabet, Amazon, and others have all raised capital expenditure forecasts to fund data centers, chips, and AI research. The combined investment is reshaping the technology industry’s cost structure.
Key dynamics of the current AI boom:
The risk of this synchronized spending is collective overcapacity. If AI service demand grows at a slower pace than anticipated, major technology companies could face underutilized assets and elevated expense structures for years.
Meta’s exposure to this risk is significant because of its large user base and heavy compute needs. The company’s decisions over the next year will influence both its own trajectory and the broader AI infrastructure market.
Investors and technology professionals should monitor several signals to gauge whether Meta’s AI spending will pay off:
Meta’s strong balance sheet provides resilience. With quarterly net income above $15 billion, the company can fund its AI ambitions without taking on substantial debt. But sustained high capex will test investor patience.
Meta’s Q3 2024 earnings illustrate a company caught between strong operational performance and the heavy costs of an ambitious AI strategy. Meta shares fell after the report, not because the quarter was weak, but because the forward-looking AI spending plan raised legitimate concerns about returns.
Zuckerberg’s “build before needed” philosophy is a bold bet on AI as a transformative platform. History shows Meta has occasionally benefited from such aggressive investments, but the scale of today’s AI spending dwarfs prior initiatives.
The takeaway for technology professionals and investors is to track Meta’s progress on tangible metrics: AI product adoption, enterprise revenue, ad efficiency, and margins. If those numbers improve, the current skepticism may give way to recognition that Meta’s AI bet was prescient. If they do not, the company faces a prolonged period of escalating costs without clear proof of return. The next few quarters will be decisive.
Meta's stock fell in after-hours trading despite strong Q3 results because investors were concerned about the company's surging AI spending. Meta raised its 2024 capital expenditure forecast to $38–40 billion and warned that 2025 capex would be "significantly higher." The market worries that these massive AI infrastructure costs may not generate clear revenue returns soon enough to justify the spending.
Meta's capital expenditures cover the physical infrastructure needed to support AI, including data centers, advanced graphics processing units (GPUs), network equipment, and energy systems. These investments are meant to expand Meta's AI compute capacity for tasks like training large models and running AI-enhanced ad targeting. The spending has been accelerating sharply, which is why it is drawing investor scrutiny.
Meta's planned capex of $38–40 billion for 2024 and even higher spending in 2025 places it among the biggest AI infrastructure investors, alongside Microsoft, Google, and Amazon. However, unlike Microsoft and Google, Meta's AI revenue streams are less clearly defined. Its enterprise AI offerings are still early stage, while its core advertising business relies on AI improvements to justify the costs.
Meta expects AI to improve ad targeting, efficiency, and effectiveness, which could boost ad revenue and help offset infrastructure costs. The risk is that the cost of AI infrastructure grows faster than the ad performance gains it delivers. If AI-driven ad improvements don't generate measurable profit quickly, Meta's margins could be squeezed.
Investors should look for concrete AI revenue disclosures, such as contributions from AI-powered advertising tools, business messaging, and enterprise products. They should also monitor capex growth versus revenue growth, as well as any signs of AI-driven margin expansion. Clear guidance about when AI investments will become profitable will be key to easing market concerns.
