The five largest US cloud providers are on track to spend $800 billion on AI infrastructure this year, representing a 94% increase over 2025 spending. Goldman Sachs strategists project that Amazon, Alphabet, Microsoft, Meta and Oracle will boost these capital expenditures by another 54% to $1.2 trillion in 2027, exceeding the Wall Street consensus of $1.1 trillion. The bank expects total lapex to reach $1.4 trillion in 2028, though the annual growth rate will slow to 12% by that time.
The massive spending scale creates significant monetization pressure, as the hyperscalers need roughly $300 billion in annual AI revenue to break even. While AI investment drives nearly half of the S&P 500's earnings per share growth in 2026, Goldman warns it could become a marginal earnings drag by 2028 due to mounting equipment depreciation. In a worst-case scenario where return on invested capital is zero, the firms would still require $920 billion annually to cover running costs and depreciation.
Key sources
- SOURCE@deitaone“boost AI infrastructure spending 54% to $1.2 trillion in 2027, above Wall Street estimates”x.com
- SUPPORT@firstadopter“hyperscaler capex growth to ease from nearly 100% this year to 54% in 2027 and to 12% in 2028”x.com
- SUPPORT@zerohedge“worst case scenario where hyperscaler ROIC on capex is 0... they still need to spend $920BN every year”x.com
- SUPPORT@rohanpaul_ai“Almost half of S&P 500 growth in EPS in 2026 comes from AI investment”x.com
- SUPPORT@zerohedge“GPU rental prices for all but the latest chips are now rolling over and token prices are at all time lows”x.com
- SUPPORT@rohanpaul_ai“The top 10% of customers (in AI) account for 99.5% of model-serving spend”x.com
- SUPPORT@stocksavvyshay“need ~$300B of annual AI revenue to break even”x.com
- SOURCEmarketbrief.now