May 2026, Goldman Sachs in a research report said AI is currently raising U.S. consumer prices through three main channels, becoming an important driver of short-term inflation. This view quickly attracted market attention. Although in the long run AI may bring disinflation through productivity gains, at the stages of infrastructure construction, software pricing and energy consumption, AI is imposing tangible cost pressure on consumers.

Goldman Sachs economists Manuel Abecasis and others stated in the report that AI-related price pressures have raised the U.S. core PCE inflation rate by about 0.3 percentage point over the past year and core CPI by about 0.1 percentage point. They expect similar effects to persist over the next year, while productivity gains have not yet fully materialized.

Three key channels through which AI raises prices

First channel: AI infrastructure demand pushes up electronic component prices

AI training and inference have driven surging demand for data centers, servers and chips, directly leading to higher prices for key electronic components such as memory and storage batteries. These rising costs are being passed through to consumer electronics. Goldman Sachs noted this has already pushed up computer accessory prices and expects it will further raise end prices for smartphones and personal computers in the coming months. Global semiconductor supply-chain tightness combined with the AI compute race amplifies this effect.

Second channel: software vendors raise prices citing AI features

Software companies are increasingly using AI features as justification for price increases. Microsoft raised Microsoft 365 subscription prices, and companies such as Adobe, Duolingo and Intuit have also increased fees after launching AI-enhanced tools. Goldman Sachs says this "AI premium" has exerted noticeable upward pressure on software prices over the past two years, meaning consumers pay more for new features. Such pricing is often treated in inflation statistics as a pure price increase rather than a quality improvement.

Third channel: data-center power demand raises electricity prices

AI data centers consume large amounts of electricity and are significantly raising power prices in some U.S. regions. Goldman Sachs expects this factor to contribute 0.1 to 0.2 percentage point to headline PCE inflation over the next few years. Higher household power bills will also indirectly raise operating costs in sectors such as healthcare, food and manufacturing, ultimately passing through to final goods and services prices. Some forecasts indicate consumer electricity prices could rise by about 6% in 2026–2027.

Far-reaching impacts on the economy, consumers and markets

1. Heavier burden on consumers

Lower- and middle-income households are hit most directly. Rising everyday expenditures on electronics, software subscriptions and power bills compress disposable income and, against the backdrop of still-elevated interest rates, could further dampen consumption. This poses a short-term conflict with the Federal Reserve's inflation-control objectives and may delay a soft landing for the economy.

2. More complex inflation dynamics

AI's short-term inflationary effects challenge market optimism about "automatic AI-driven disinflation." Institutions like Goldman Sachs believe productivity gains will take time to spread across the broader economy. Until then, the AI investment boom itself is a new inflationary variable. This could affect the Fed's policy path, delay rate-cut expectations and increase policy uncertainty.

3. Reshaping industry structure and investment patterns

AI infrastructure-related firms (such as chipmakers, power and data-center companies) are clear short-term beneficiaries, but higher costs for end-co