AI's race to transform the world before the money runs out
NVDA•Global data-center spending could exceed $30 trillion by 2050, while Anthropic plans to spend $518 billion in coming years, according to figures cited in the article. Economists and studies question whether productivity gains and new revenue will arrive quickly enough to support AI investment.
1. Spending and revenue gap
A PwC projection puts cumulative global data-center spending above $30 trillion by 2050. Anthropic plans to spend $518 billion in coming years, more than 100 times its 2025 revenue, according to its IPO prospectus. Bain said US hyperscalers and other AI companies need more than $4.2 trillion in new revenue over five years to fund the buildout.
2. Productivity remains uncertain
JPMorgan said broad-based US productivity gains from AI “remain elusive.” Bain said gains in existing markets would not justify current spending and that entirely new markets must emerge. Columbia Business School economist Stijn Van Nieuwerburgh estimated US AI investment could reach about $9 trillion from 2025 to 2032, and said the sector would need about $3.55 trillion in annual revenue by 2032 to earn a 10% return.
3. Technology takes time
Economist Diane Coyle said the productivity impact of past revolutionary technologies usually took about 10 to 50 years to emerge. Anthropic modeled annual growth in 2030 ranging from 2.4% under a modest AI impact scenario to 15.4% in an extreme scenario, against a 2% non-AI baseline. Stanford researchers said employment of workers aged 22 to 25 in AI-exposed industries was 19% lower than in jobs AI was less able to replicate.
