AI spending is outrunning revenue as hyperscalers chase returns

AI companies are hitting a tougher phase in the boom as spending on chips, data centers and other infrastructure continues to run ahead of revenue. Axios reports that two Stanford economists estimate a nearly $1 trillion gap between hyperscaler AI spending and AI revenue since 2024, and argue those companies would need to triple or quadruple AI revenue next year and keep that pace for a decade to justify the investment. Axios also cites Goldman Sachs as finding current AI revenue still below break-even on capital spending. At Anthropic, Reuters reported last year’s $4.6 billion in revenue came with an operating loss nearly twice as large, even as Axios says the company’s annualized revenue run rate topped $65 billion in the second quarter.