According to Business Insider, Meta reported a strong Q4 2025 with revenue of $59.89 billion, beating Wall Street forecasts and causing its stock to jump up to 10% after hours. The company’s ad revenue was $58.14 billion for the quarter, up 24% year-over-year. However, CEO Mark Zuckerberg signaled a massive spending shift, with capital expenditures for 2026 projected to hit $115-$135 billion, nearly double the $72 billion spent in 2025, to fund its AI ambitions. CFO Susan Li warned total 2026 expenses could reach $169 billion, driven by infrastructure and hiring technical AI talent. Meanwhile, the Reality Labs division posted a record quarterly loss of $6.02 billion.
The 2026 Money Furnace is Lit
Look, the numbers here are just staggering. Meta is basically telling investors to buckle up for a year where they might spend almost as much on AI infrastructure as some countries spend on their military. The jump from $72 billion to up to $135 billion in CapEx isn’t just an increase; it’s a statement of absolute urgency. Zuckerberg clearly believes that if Meta doesn’t own the foundational AI models, it risks becoming a tenant in the next era of tech, and he’s willing to burn mountains of cash to avoid that fate. The real question is, how long will Wall Street’s patience last with this level of investment before demanding clearer paths to new revenue? The ad machine is humming now, but it has to fund this entire war.
Ads: The Golden Goose Funding the Sci-Fi Lab
Here’s the thing: all this futuristic talk rests entirely on the shoulders of the very present, very lucrative ad business. Zuckerberg openly admitted ads will be the main growth driver “for the next couple of years.” It’s a classic tech company move—use the cash cow to fund the moonshot. But Meta’s pitch is clever: they’re arguing the AI spend isn’t just for future products; it’s already making the ad cow produce more golden eggs. Tools for auto-generating video ads and better measurement are hitting multi-billion-dollar run rates. So it’s not a pure gamble. They’re trying to show that the AI investment has an immediate, measurable return on the core business, which is a smarter narrative than just asking for faith in the metaverse.
Reality Labs: The Pivot is Real
The “metaverse” as a standalone destination? Basically, that vision is on life support. Reality Labs lost over $19 billion in 2025, and Meta expects similar losses in 2026. That’s a brutal drag. But the strategy has visibly shifted. Now, Zuckerberg is talking about AI-powered smart glasses and bringing “immersive 3D” experiences—think games created from a single prompt—directly into the Facebook and Instagram feeds. Horizon Worlds isn’t dead, but its future is as a feature inside apps used by billions, not a separate VR universe. It’s a pragmatic, if less revolutionary, path. They’re admitting that getting people to put on a headset is hard, but getting them to tap on a 3D object in their scroll is easy. It might just work.
The AI Efficiency Paradox
This is the most fascinating internal bet. Zuckerberg says AI will let them flatten teams and that projects needing big teams can now be done by a single “very talented person.” Li claims a 30% increase in output per engineer. So they’re spending unprecedented amounts on compute to… hopefully need fewer people? Or at least get vastly more from them. It’s the Silicon Valley dream: automate the grind, elevate the creators. But it creates a strange paradox. They’re adding huge costs (infrastructure, AI talent salaries) to reduce other costs (hopefully, overall headcount growth). The success of this entire spending spree might hinge on whether those internal productivity gains materialize at scale. If they do, Meta could emerge leaner and more powerful. If not, well, that’s a very expensive experiment.
