Goldman Sachs Sees $1.4T AI Infrastructure Spend by 2027

Goldman Sachs expects U.S. hyperscalers to spend $1.4 trillion in capital by 2027, as artificial intelligence shifts from experimentation to implementation and consumer AI agents open a new phase of growth.

This estimate is above Wall Street consensus, with Goldman Sachs predicting the AI infrastructure spending cycle will remain elevated through 2027.

Eric Sheridan, business unit leader at Goldman Sachs Research, noted that companies are moving beyond experimenting with AI to implementing it in both internal operations and customer-facing interactions.

The growth of consumer AI agents is expected to be driven by tasks such as shopping, travel bookings, and calendar management, which will be monetized through advertising and subscriptions.