Mark Zuckerberg Bets $600 Billion on AI Future: “Better to Overbuild Now Than Regret Later”

Meta CEO Mark Zuckerberg is making one of the boldest bets in tech history — a staggering $600 billion investment in artificial intelligence infrastructure by 2028. In his latest remarks on the Access podcast, Zuckerberg made it clear that Meta intends to lead the AI revolution, even if it means facing potential risks from an overheated market.

Speaking candidly, Zuckerberg acknowledged that the ongoing AI boom might be a bubble but argued that inaction poses a greater danger. “If we end up misspending a couple of hundred billion dollars, I think that is going to be very unfortunate,” he said. “But the risk is higher on the other side.”

The Meta chief emphasized that the company’s strategy is rooted in long-term vision rather than short-term caution. “The risk, at least for a company like Meta, is probably in not being aggressive enough rather than being somewhat too aggressive,” Zuckerberg added, underscoring his conviction that building big today will pay off when artificial superintelligence (ASI) becomes a reality.

Zuckerberg drew parallels between today’s AI frenzy and previous eras of tech over-enthusiasm, noting that even when those markets crashed, they left behind valuable infrastructure that fueled future innovation. He believes the same logic applies to AI — even if there’s a temporary bubble, the infrastructure Meta builds will power the next technological leap.

Earlier this month, Meta confirmed plans to pour at least $600 billion into U.S.-based data centers and infrastructure projects by 2028. The company’s CFO, Susan Li, clarified that this figure includes all domestic data-center expansions, workforce growth, and related U.S. investments. The announcement stunned analysts and reinforced Meta’s commitment to staying ahead of rivals like OpenAI and Anthropic.

While competitors rely heavily on investors to fund their computing resources, Meta enjoys a unique advantage — vast financial independence. “We’re not at risk of going out of business,” Zuckerberg said, in a pointed reference to AI startups dependent on external funding. He cautioned that market downturns could cripple smaller labs by cutting off access to costly infrastructure. “It might be a different situation if you’re in one of their shoes,” he added.

Meta’s latest move also includes the creation of a dedicated “superintelligence lab”, uniting top AI researchers under one roof. This elite team operates with minimal hierarchy and no fixed deadlines — a deliberate attempt to foster innovation without corporate constraints.

At the core of Meta’s strategy is a new metric: “compute per researcher.” Zuckerberg believes that giving scientists unrestricted access to computing power will unlock the next wave of breakthroughs. To that end, Meta is investing billions in GPUs and custom AI infrastructure designed to handle the immense demands of future AI systems.

By choosing to “overbuild now,” Zuckerberg is staking Meta’s future on the belief that artificial superintelligence will arrive sooner than most expect — and that only those with the courage (and capital) to invest massively today will dominate tomorrow’s AI landscape.