Mark Zuckerberg predicts that billions of people will have personal AI agents in five years

By Aoife Gallagher · Reporting from Dublin ·

The scale of the pronouncements coming out of Silicon Valley these days is breathtaking—a dizzying cascade of "super-intelligence," "personalized agents," and inevitable global transformation.

The Illusion of Universal Super-Intelligence

The scale of the pronouncements coming out of Silicon Valley these days is breathtaking—a dizzying cascade of "super-intelligence," "personalized agents," and inevitable global transformation. On July 29, 2026, Mark Zuckerberg laid down a prediction so vast it borders on prophecy: that within five years, billions will rely on personal AI agents working on their behalf, managing everything from finances to relationships. The narrative is one of seamless convenience, where every user has an infinitely patient, always-on digital assistant. But when you peel back the layers of this glossy promise—the kind of future Zuckerberg painted in his op-ed for The Guardian, suggesting that universal access to a "super-intelligent lawyer" would lead to fairer justice—what emerges is less a vision of utopia and more an industrial blueprint for total market capture. Meta, meanwhile, continues its complex dance between generating cash from digital ads (which still account for 98% of their revenue, according to CNBC) while simultaneously pouring billions into the next frontier. It’s a staggering act of engineering confidence built on infrastructure that is barely visible to the average person.

The Weight of Infrastructure and Lost Capital

The numbers themselves tell a more complicated story than the breathless pronouncements allow. While Meta reported total revenue of $60.8bn, beating analyst predictions, the financial picture is far from stable or simple. TechCrunch noted that while they are making money on ads, Reality Labs hemorrhaged another $4.62 billion in the latest period. The running total losses since 2021 now approach $88 billion—a figure so gargantuan it demands serious scrutiny. Zuckerberg himself acknowledged this dilemma: “A common trade-off that we need to make is around how much do you monetize something today versus develop future assets,” he stated on the earnings call cited by CNBC. This isn't merely an investment; it’s a bet of existential proportions, one where current profits are being systematically sacrificed at the altar of computational capacity. The company needs compute—and they need to convince investors that this massive expenditure is not just cost, but necessary pre-condition for future revenue streams.

From Horsepower to Habit: A Familiar Pattern

What Zuckerberg and Meta are selling isn't a piece of software; it’s the next foundational convenience—a new utility layer that will redefine daily human movement and interaction in the digital sphere. This entire performance echoes, with unnerving clarity, The automobile industry's rise. The mechanism is identical: industrializing a foundational convenience to create an entirely self-sustaining economic apparatus. In the United States, the shift from horse-drawn carriages to mass-produced vehicles didn't just change how people moved; it created an entire new ecosystem—gas stations, highways, roadside motels, and the whole supporting infrastructure that defined modern life. This precedent supports my judgment completely. Just as early automotive giants controlled not just the car itself but the network of roads required for it, Meta is positioning WhatsApp and Messenger—the messaging surfaces where these agents will live—as the indispensable infrastructure layer. The control over the format of interaction, rather than merely the content, is the true prize.

The promise of personalized super-intelligence suggests a decentralized, equitable future. But history shows that when such foundational convenience is industrialized and controlled by private entities—regardless of how benevolent their stated intent—the first beneficiaries are always those who control the rails. The small nations, the periphery, the ones who rely on established rules for survival, must watch this closely. Meta’s bet isn't just on AI agents; it's on making itself the invisible infrastructure that every single digital transaction and personal decision must pass through.

The inevitable trajectory is clear: these powerful agents will not be a tool of universal empowerment but rather an immensely profitable new choke point, solidifying control over human attention and action within Meta’s walled garden.

Sources

  1. CNBC: Zuckerberg lays out Meta's AI capacity dilemma: What to sell vs. what to keep
  2. The Guardian: Meta misses earnings forecasts after Zuckerberg media push to promote AI
  3. techcrunch.com: Mark Zuckerberg predicts that billions of people will have personal AI agents in five years | TechCrunch
  4. pymnts.com: Zuckerberg Says Billions Will Rely on Personal AI Agents
  5. aitoolly.com: Zuckerberg Predicts Billions of AI Agents in 5 Years | AIToolly