Investing Ahead: Zuckerberg Admits Meta May Not Use All AI Funds Soon

Meta plans to spend $72 billion in 2025 on building infrastructure for artificial intelligence — one of the largest single-year technology investments in history. However, CEO Mark Zuckerberg acknowledged during a recent meeting with investors that not all of this funding may be fully utilized in the near term.

Following his statement, Meta’s shares dropped by around 8%, reflecting investor uncertainty about the company’s short-term strategy. Despite this, Zuckerberg emphasized that these are “investments for the future”, designed to prepare Meta for the potential rapid emergence of artificial superintelligence (ASI).

According to him, the AI infrastructure being developed will ensure that Meta is technologically prepared if ASI arrives sooner than expected. If the breakthrough takes longer, the company plans to use the excess computing power to enhance its core platforms — Facebook and Instagram, which still generate over 97% of Meta’s revenue through advertising.

Industry analysts estimate that global corporate spending on AI infrastructure could reach $400 billion by 2027, with Meta, Google, and Microsoft leading the race. While Meta’s current AI tools — such as content recommendation algorithms and automated ad systems — have improved efficiency, they haven’t yet revolutionized user experience.

Zuckerberg stated that AI expenses will grow even further in 2026, as Meta aims to maintain technological leadership. The CEO underlined that these investments are “a bet on readiness”, ensuring Meta won’t fall behind when AI capabilities accelerate dramatically.

He added that the worst-case scenario would be to temporarily slow infrastructure expansion until existing capacities are fully utilized — a trade-off he considers acceptable for long-term dominance in the AI race.