Goldman Sachs sees annual AI-related spending rising toward $1.4 trillion by end-2027

Goldman Sachs is forecasting a sharp rise in global spending tied to artificial intelligence, projecting that annualized spending on AI-related hardware and intellectual property will climb to $1.4 trillion by the end of 2027, according to reporting published Wednesday.

The estimate focuses on two broad categories. “Hardware” refers to the physical systems used to run AI, such as data center infrastructure and computing equipment. “Intellectual property” refers to rights and assets companies license or buy to use AI technology, including software and related IP protections.

For markets, the forecast matters because it frames AI investment not as a series of one-off pilots, but as a multi-year build cycle that can feed demand for semiconductors, networking, data center equipment, and enterprise software. It also suggests companies may treat AI costs as recurring operating or investment expenses rather than purely experimental spending.

Goldman’s projection arrives as investors continue to reassess which parts of the AI supply chain can translate demand into sustained revenue. The bank’s “annualized” wording implies the spending level would be measured as if run at a full-year rate by the end of 2027, rather than only looking at incremental changes quarter to quarter.

The forecast also points to a wider view of AI spending than computing alone. By including intellectual property in the spending basket, the estimate recognizes that adopting AI often requires licensing and purchasing software layers and technology rights, not just buying servers or chips.

Still, the company did not provide, in the published reporting, additional operational detail such as the model assumptions behind the $1.4 trillion figure, how it breaks down across geographies or industries, or which cost components are driving the growth most sharply.

There is also no disclosed timetable for how quickly the spending ramp is expected to occur, beyond the endpoint at the end of 2027. Without those specifics, market participants may need follow-up from Goldman, including any full report publication, to understand how the forecast might translate into sector-level expectations.

Looking ahead, investors will likely watch for whether other banks, research houses, and industry groups converge on similar ranges, and whether reported capex plans from technology and telecom companies align with an AI spending run-rate approaching the forecast’s magnitude by late-2027.