Meta Platforms is set to report its fourth-quarter and full-year 2025 results after the market close on Wednesday, but the most anticipated number won't be revenue, earnings, or even advertising growth. What investors truly want to know is how much CEO Mark Zuckerberg plans to spend on artificial intelligence in 2026—and whether that figure will cross into unprecedented territory.

The $113 Billion Question

Bank of America analyst Justin Post has set the bar for expectations, projecting Meta's 2026 capital expenditures could reach $113 billion—a 58% increase from the $70-72 billion the company guided for 2025. When combined with operating expenses, Post estimates Meta's total spending for 2026 could range between $153 billion and $160 billion.

To put that in perspective: $113 billion in capital expenditures would exceed the entire annual revenue of all but a handful of companies globally. It would dwarf the GDP of more than 100 nations. And it would represent the largest single-year infrastructure investment by any technology company in history.

"Forward-looking guidance for 2026 capital expenditures will be the primary catalyst for stock price volatility post-earnings. With management having previously signaled that 2026 spending will be 'notably larger' than 2025, market fears have coalesced around the potential for CapEx to breach the $100 billion threshold."

— Market analysis

What's Driving the Spending Surge

Meta's AI infrastructure buildout spans multiple dimensions, each requiring massive investment:

Data Centers

The company is constructing what it calls the "world's largest AI data center" in Louisiana, along with major facilities in other U.S. locations. These facilities require not just buildings but specialized power systems, cooling infrastructure, and networking capabilities designed for AI workloads.

Custom Hardware

Meta has developed its own AI training chips and continues to purchase enormous quantities of Nvidia GPUs. The company reportedly ordered hundreds of thousands of Nvidia's latest Blackwell processors, each costing tens of thousands of dollars.

Model Training

Training frontier AI models requires computational resources that scale exponentially. Each generation of Llama models demands more computing power than the last, and Meta is racing to keep pace with—or surpass—OpenAI and Google.

The Revenue Side of the Equation

While the spending numbers grab headlines, Meta's Q4 results should demonstrate that the company can afford its ambitions. Wall Street consensus projects:

  • Revenue: $58.45 billion, representing 21% year-over-year growth
  • Advertising revenue: $56.85 billion, up 21.5% year-over-year
  • Earnings per share: $8.16, a modest 2% increase

The relatively slow earnings growth despite strong revenue reflects the impact of AI investments on profitability. Meta is essentially choosing to reinvest its advertising windfall into infrastructure rather than flow it through to the bottom line.

Meta AI's Massive Scale

The company's AI products have achieved remarkable reach. Meta AI now has over 1 billion active users across Facebook, Instagram, WhatsApp, and Messenger. This gives Meta something rivals lack: immediate distribution for AI features to a massive existing user base.

Investor Concerns and Debates

The central question dividing Wall Street is whether Meta's AI spending represents visionary investment or reckless capital destruction. Both camps have compelling arguments:

The Bull Case

AI-powered improvements to ad targeting have already driven measurable revenue gains. Better recommendations keep users engaged longer. New AI features could unlock entirely new revenue streams. And if AI truly transforms the economy, being a leader matters enormously.

The Bear Case

Returns on AI investment remain largely speculative. Spending $113 billion requires assumptions about monetization that haven't been proven at scale. And the competitive dynamics of AI could mean today's investments become obsolete faster than traditional infrastructure.

Reality Labs: The Other Money Pit

Lost in the AI spending discussion is Meta's continued investment in virtual and augmented reality through Reality Labs. The division has lost over $50 billion since its formation and continues to bleed cash, though at a slower rate as Meta has pulled back some metaverse ambitions in favor of AI.

Analysts expect Reality Labs losses of approximately $4-5 billion in Q4, continuing the drain on overall profitability. The division's future remains uncertain, though the Quest headset line continues to dominate the VR market.

Stock Performance and Valuation

Meta shares have gained about 14% year-to-date heading into earnings, though they've underperformed the broader Nasdaq. The stock trades at roughly 22 times forward earnings—expensive by historical Meta standards but cheap compared to other Magnificent Seven names.

Options market pricing implies traders expect an 8% move in either direction following the earnings release, reflecting the uncertainty around spending guidance. The stock has a strong consensus rating from analysts, with 37 buy recommendations, 6 holds, and just 1 sell.

What to Watch Tonight

Beyond the 2026 CapEx guidance, investors should focus on several key metrics:

  • Daily active users: Any signs of engagement plateauing would concern investors
  • Average revenue per user: The key indicator of advertising pricing power
  • AI monetization commentary: How is AI actually contributing to revenue today?
  • Regulatory update: Any developments on FTC antitrust matters or global regulatory challenges

Meta's Q4 report will provide the clearest window yet into how the tech industry's largest AI bet is progressing—and whether the scale of investment Zuckerberg envisions makes sense. For better or worse, the answer will help determine not just Meta's trajectory but expectations for technology investment across the sector.