According to PYMNTS.com, on Meta’s Q4 2025 earnings call, Chairman Mark Zuckerberg told analysts their questions would be “somewhat unfulfilling” as the company is six months into rebuilding its AI efforts, with more to share in the coming months. He pledged a massive capital expenditure increase, targeting $115 to $135 billion by 2026—nearly double 2025’s $72 billion—to fund new AI data centers and work on a “superintelligent” model. This spending could outpace Google’s. The push is backed by a roaring ad engine: Q4 revenue was $59.89 billion, up 24%, with profit at $22.76 billion. Zuckerberg outlined a vision for “personal superintelligence” and “agentic shopping” tools, naming smart glasses as the “ultimate incarnation” of an AI assistant, with sales of them more than tripling last year.
The Staggering Scale of the Bet
Look, the numbers here are just wild. We’re talking about a company planning to potentially spend $135 billion in a single year, 2026. That’s not a typo. For context, that’s more than the entire market cap of some major corporations. It’s a declaration of total war in the AI arms race. Zuckerberg is basically saying, “We’re not playing for incremental gains; we’re betting the company on becoming an AI-first entity.” The fact that CFO Susan Li had to repeatedly reassure everyone that operating income will still be higher in 2026 tells you everything. They know how insane this sounds to Wall Street. They’re preemptively trying to calm the nerves they’re deliberately fraying.
The Vision: Agentic Everything and Glasses
So what’s the grand plan for all this cash? Zuckerberg’s talking about “personal superintelligence”—AI that knows your history, your goals, your weird obsessions. It’s about folding large language models into the recommendation engine so it doesn’t just suggest a reel, but understands you’re trying to, I don’t know, build a deck and finds you the perfect tutorial and the lumber supplier. The “agentic shopping” angle is obvious: supercharge an already terrifyingly effective ad system. But here’s the really interesting part: he’s pegging the ultimate form factor to smart glasses. He’s comparing this moment to flip phones to smartphones. That’s a huge, risky narrative to hang your hat on. Reality Labs revenue was still under $1 billion last quarter. Tripling a small number is still… a relatively small number. This is a decade-long bet being funded by today’s ad clicks.
The Massive Hidden Risk
And this is where the skepticism has to kick in. Meta has a history of pouring billions into grand visions that… well, take a long time to materialize, if they ever do. The Metaverse, anyone? Now, they’re pivoting hard to AI, which is the hot thing, but they’re doing it at a scale that could cripple a less robust company. The risk isn’t just that the AI doesn’t pan out. It’s that this spending becomes a permanent new baseline. Data centers and chip contracts aren’t one-off expenses. They create a forever-higher cost of doing business. What happens if the ad market hiccups? What if regulators finally break the algo? This level of capex burns your bridges. There’s no easy way back. You’re all-in.
Can The Ad Machine Fund It All?
Basically, the entire thesis rests on one thing: that the legacy social media ad business remains an unstoppable cash geyser. Right now, it is. The AI improvements in ad targeting are already feeding that beast, making it stronger. It’s a virtuous cycle—for now. But you have to ask: is Zuckerberg building the next generation of Meta, or is he just building a fantastically expensive tool to optimize the last generation? The “personal superintelligence” that helps you “improve your life” sounds great, but in practice, it will be measured by how well it sells shoes and keeps you scrolling. The gamble is that by spending more than anyone else, they can force a new paradigm—one where the glasses and the agents are so useful we forget they’re also the ultimate advertising and data collection platform. It’s a bold move. But at $135 billion, “bold” might just be another word for “terrifying.”
