Earnings Look Great, But Meta Is Still in the 'Pending Zone'

Earnings Beat Expectations, but the Market Isn't Buying It

Meta delivered a near-flawless report card on paper. Revenue and profit for the latest quarter both exceeded Wall Street analyst expectations, the advertising business remains a money-printing machine, user growth is steady, and virtually every core metric is free of obvious weaknesses. Yet immediately after the earnings release, the market took a fright — investors voted with their feet and delivered a starkly different verdict.

This scene is hardly a first. Over the past several quarters, Meta's earnings narrative has settled into a peculiar pattern: the better the results, the more scrutiny the market applies to its AI spending. What Wall Street truly cares about is no longer "how much Meta is earning now," but "whether the money it's spending will ever pay off."

The AI Gamble: Staggering Spending

Meta's commitment to AI ranks among the most aggressive in the entire tech industry. Mark Zuckerberg has publicly stated on multiple occasions that AI is Meta's most important strategic direction for the next decade, and the company will spare no expense in building AI infrastructure.

Capital expenditure data tells the story: spending on data center construction, GPU procurement, and AI R&D continues to climb, with full-year capex guidance reaching levels that made analysts gasp. The money is flowing into several key areas:

• Massive GPU cluster buildouts: Providing the compute power to train and run inference on the Llama family of large language models
• AI infrastructure expansion: Constructing and expanding data centers worldwide
• AI product development: Embedding AI capabilities into the advertising system, content recommendations, Meta AI assistant, and other product lines
• Open-source ecosystem investment: Continuously iterating on Llama models and maintaining influence in the open-source community

This "invest heavily first, seek returns later" strategy is not without precedent. Meta went through a similar period of pain when it bet on the mobile transition years ago, and ultimately proved itself right. But this time the stakes of the AI wager far exceed that era, and competitors are placing equally aggressive bets.

The Commercialization Fog: Money Spent, but Where Are the Returns?

The core question unsettling the market is: when and how will Meta's AI investments produce quantifiable returns?

Currently, Meta's AI commercialization roadmap can be roughly broken down into several threads:

First, AI upgrades to the advertising system. This is the clearest and most market-validated path. Meta is using AI to optimize ad targeting precision and efficiency, and AI-driven advertising tools such as Advantage+ are already creating measurable value for advertisers. The strong advertising revenue in the earnings report is partly attributable to this. The problem, however, is the enormous order-of-magnitude gap between the incremental revenue from ad optimization and the colossal investment in AI infrastructure.

Second, user growth for the Meta AI assistant. Zuckerberg has repeatedly emphasized that Meta AI's user base is growing rapidly, with the goal of becoming the world's most widely used AI assistant. But there is a chasm between "lots of users" and "making money." Meta AI currently functions primarily as a free feature embedded in WhatsApp, Instagram, and Facebook, and has yet to develop an independent business model.

Third, the strategic value of the Llama open-source ecosystem. The open-source strategy for the Llama model family has earned Meta widespread goodwill from the developer community and significant industry influence, but open source itself does not directly generate revenue. Its strategic logic is more about "reducing the industry's dependence on competitors' closed-source models through open source, thereby indirectly protecting Meta's core business" — a story that is rational but difficult to validate with financial data.

Fourth, AI-driven content generation and interaction. Meta is exploring new formats such as AI-generated content and AI character interactions, aiming to increase user dwell time and engagement on its platforms. These experiments are still in early stages, and their commercial prospects remain unclear.

Why Did the Market Take Fright?

With the above context in mind, the market's reaction is not hard to explain. Investors face a classic tug-of-war between certainty and uncertainty:

What is certain: Meta's core business is very healthy right now — advertising revenue is robust, profit margins are impressive, and cash flow is ample. Judging purely by the present, this is a company with exceptional profitability.

What is uncertain: Meta is pouring vast profits into an AI track with an unknown return cycle, an unsettled competitive landscape, and an unproven business model. Even more concerning, there is no sign of this spending slowing down in the near term — management has repeatedly emphasized that "investment will continue to increase."

For Wall Street, this uncertainty means a larger discount must be applied in valuation models. When a company's free cash flow is being eroded by massive capital expenditure, future shareholder returns will be diluted no matter how good current profits look.

Moreover, there is a deeper layer of anxiety in the market: if AI commercialization falls short of expectations, does Meta have the willingness and ability to "hit the brakes"? Recalling how Meta's obsession with the metaverse in 2022 led to a stock price that was cut in half, investors have fresh memories of Zuckerberg's "spare no expense" strategic style.

Peer Comparison: Meta Is Not Alone, but Its Situation Is Unique

It is worth noting that Meta is not the only tech giant being questioned by the market over AI spending. Google, Microsoft, and Amazon all face similar "AI capex anxiety." But Meta's situation has its own unique characteristics:

• Unlike Microsoft and Google, Meta has no cloud computing business that can directly convert AI compute into enterprise customer revenue
• Unlike Amazon, Meta has no ready-made channel like AWS to distribute AI services
• Meta's AI monetization is heavily dependent on its social advertising ecosystem, making its path relatively narrow

This means Meta needs to find killer AI application scenarios within the "advertising + social" framework, rather than simply packaging and selling AI capabilities the way cloud providers can. This is both a constraint and a potential opportunity — if Meta can break through in consumer-grade AI applications, its social platforms reaching billions of users will be an unparalleled distribution advantage.

Zuckerberg's Confidence vs. the Market's Patience

Zuckerberg's remarks on the earnings call were as resolute as ever. He emphasized that AI will reshape every one of Meta's product lines, that Llama model capabilities are improving rapidly, and that Meta AI user growth has exceeded internal expectations. His core argument can be summarized as: "Now is the investment phase; returns will materialize gradually over the next few years."

This argument is not without merit. AI technology is indeed advancing rapidly, and the results of applying AI to Meta's advertising system are genuinely impressive. But the market's patience has its limits. Investors need to see more concrete milestones — not just user number growth, but tangible revenue contributions and margin improvements.

The current situation can be captured with an analogy: Meta is building a highway to the future. The roadbed is laid and construction is progressing quickly, but whether there is truly a "gold mine" at the destination is something no one can say with complete certainty.

Outlook: When Will the 'Pending Zone' Deliver a Verdict?

In the short term, the return on Meta's AI investment will remain in "pending" status. The following signals are worth watching closely:

1. Penetration rate and incremental contribution of advertising AI tools: If tools like Advantage+ can continue to drive growth in ad ARPU (Average Revenue Per User), that will be the most direct positive signal
2. Commercialization progress of Meta AI: Whether paid features, enterprise editions, or deep e-commerce integrations will be introduced
3. Competitiveness of Llama models: Whether they can maintain a technological lead — or at least keep pace — in competition with GPT, Claude, and Gemini
4. An inflection point in capex growth: What the market most wants to see is a signal that AI investment is shifting from "accelerated expansion" to "efficiency optimization"

In the long run, Meta's success or failure in the AI race may hinge on a fundamental question: What is the ultimate form of consumer-grade AI? If AI assistants become gateway applications in people's daily lives, then Meta, with the world's largest social network, will hold a natural advantage. But if AI's primary value ultimately manifests in enterprise services and productivity tools, Meta's competitive position will be relatively passive.

Strong earnings can only prove that Meta's "present" is fine, but the capital markets price in the "future." Until the dust settles on this AI gamble, Meta will likely remain in the "pending zone" for some time to come. For investors, this represents both risk and potential opportunity — the key question is whether you are willing to bet on this future alongside Zuckerberg.