Goldman Sachs says AI model costs fell in three years as much as PC prices did in 15 years

Goldman Sachs data shows that the decline in AI model costs over the past three years is comparable to the cumulative drop in personal computer prices over the last 15 years. The bank said the speed of that cost compression could weigh on the profit margins of major technology companies. It also warned that the trend may challenge infrastructure investment if demand does not keep pace with expanding supply. The comparison points to how quickly economics in the AI sector are shifting, with falling costs potentially changing returns across the broader technology stack. Goldman Sachs framed the issue around the balance between supply growth and market demand rather than around a specific company or product, highlighting margin pressure and investment risk as the key areas to watch.