Meta CEO Mark Zuckerberg has acknowledged that an AI bubble is “quite possible,” drawing parallels to previous tech booms and busts. But while investors warn of overheating in the market, Zuckerberg insists the bigger danger lies in building too slowly and missing out on the arrival of artificial superintelligence.

“If we end up misspending a couple of hundred billion dollars, I think that that is going to be very unfortunate, obviously,” Zuckerberg said on the Access podcast. “But what I'd say is I actually think the risk is higher on the other side.”

The Growing Talk of an AI Bubble

The comments come as debates over an AI bubble intensify, with many recalling the dot-com collapse of 2000. Some experts fear that massive investments in artificial intelligence could outpace real-world returns, leaving companies and investors exposed. Why Everyone Is Talking About an AI Bubble - Business Insider

Meta has made some of the largest commitments in the AI industry, pledging at least $600 billion on U.S. data centers and infrastructure through 2028. The scale of the investment highlights Zuckerberg’s conviction that hesitancy could leave the company behind in the race toward artificial superintelligence.

Meta’s Multi-Billion-Dollar Gamble

Zuckerberg stressed that being “out of position” when superintelligence arrives would be a far greater setback than overspending. “The risk, at least for a company like Meta, is probably in not being aggressive enough rather than being somewhat too aggressive,” he explained.

Meta CFO Susan Li later clarified that the $600 billion pledge includes not only AI data center infrastructure, but also hiring and other investments tied to Meta’s U.S. operations.

Meta CFO Explains Mark Zuckerberg's $600 Billion White House Pledge - Business Insider

Balancing Investor Fears and Market Pressures

Despite Meta’s bold commitments, investors are voicing concerns about runaway spending and whether AI is truly capable of delivering the long-promised leap to superintelligence. Wall Street analysts have scrutinized Meta’s labor costs and raised alarms over stock-based compensation potentially diluting shareholder value.

Burning Man Festival Death: Authorities Release Key Evidence in Vadim Kruglov Homicide - USA Herald

Last month, Meta froze some recruitment efforts after previously offering massive signing bonuses to lure top AI researchers. The pullback followed growing concern that the AI arms race could drive unsustainable costs.

Meanwhile, rivals like OpenAI and Anthropic rely heavily on outside fundraising to cover massive compute expenses. Zuckerberg pointed out the difference: “We’re not at risk of going out of business. It might be a different situation if you’re in one of their shoes.”

Preparing for Superintelligence

Meta is now focusing on consolidating elite talent within its superintelligence lab, which emphasizes flexible research with no top-down deadlines. The company is also investing heavily in custom infrastructure, making “compute per researcher” a key advantage in outspending rivals.

“Meta is preparing for superintelligence by concentrating elite talent in a small, flat ‘superintelligence’ lab,” Zuckerberg explained. He added that this structure allows researchers to focus on breakthroughs without corporate time pressure. Why Zuckerberg Says Risking Billions on AI Is Worth It – Business Insider

AI Bubble Market Reality

For now, Meta’s gamble is paying off — its stock has risen nearly 40% in the last year. Still, the conversation around an AI bubble continues to dominate tech and investment circles, with fears of a dot-com-style bust lingering.

Zuckerberg remains confident. To him, the cost of missing superintelligence is far greater than the price of overspending. “What I think is going to be the most important technology that enables the newest products and innovation and value creation in history” cannot be risked, he argued.