Goldman Sachs predicts $1.4 trillion in U.S. hyperscale capital expenditures by 2027 as consumer AI agents emerge.

Huoxing Finance reports that on September 18, Goldman Sachs Research released a perspective stating that AI is transitioning from the experimental phase to implementation, with the rise of consumer-grade AI agents signaling the emergence of a platform layer. At the Communacopia + Technology Conference in San Francisco, most companies showcased case studies demonstrating this shift from experimentation to implementation. Goldman Sachs forecasts that U.S. hyperscale enterprise capital expenditures will reach $1.4 trillion by 2027, exceeding Wall Street consensus. Goldman Sachs analyst Eric Sheridan noted that consumer-grade AI agents are evolving from conversational interactions to action-oriented functions; if consumers overcome trust and security concerns, these agents could execute complex tasks such as booking tickets and reserving hotels. Monetization of these agents at scale is expected to resemble search, driven by advertising and subscriptions. AI is evolving from the infrastructure layer to the platform and application layers, with declining per-token pricing and improved utility serving as key drivers for mass adoption. While AI-related risks emerged as a primary topic of discussion at the conference, Goldman Sachs believes this will not slow infrastructure development, as demand for computing power continues to outstrip supply and most projects are already under contract. Constraints in the supply chain—such as memory chips, electricity, and land—may pose challenges, but capital expenditure levels are expected to remain high through 2027.