Generative AI’s $172 Billion Value to U.S. Consumers: What the Estimate Means

What “consumer surplus” means

Many generative-AI tools are free or priced far below what some users believe the tools are worth. Consumer surplus is an economic estimate of the difference between what people would be willing to give up to keep access and what they actually pay.

This makes consumer surplus useful for evaluating free digital services whose contribution can be invisible in company revenue or conventional economic statistics.

How the estimate was produced

Researchers conducted online choice experiments with approximately 1,400 participants in 2025 and 2,000 in early 2026. Participants were asked how much compensation they would require to give up access to all generative-AI tools for one month.

The results were annualized and extrapolated across the estimated U.S. user population. Stanford reports that U.S. adult use rose from 48% to 56%, or roughly 95 million to 115 million users. Average monthly surplus per user increased from $98 to $125, while the median rose from $3.40 to $11.40.

What the number does not mean

• It is not $172 billion of company revenue.
• It is not a measured addition to GDP.
• It is not proof that AI increased worker productivity by $172 billion.
• It is not money consumers actually received or spent.
• It depends on stated choices in experiments and on extrapolation to the broader U.S. population.

Why it matters

The estimate illustrates a central AI paradox: a technology can create enormous social and consumer value even when users pay little for it and producers struggle to capture that value as durable profit.

For business leaders and investors, the distinction matters. Rapid adoption and genuine usefulness can coexist with intense competition, falling prices, weak switching costs and pressure on margins. Creating value and capturing value are not the same economic achievement.

What to watch next

The key questions are whether the user base continues expanding, whether median value keeps rising, and whether paid products capture more of the value users currently receive. Future estimates should also be compared with observed productivity, revenue and labor-market data rather than treated as substitutes for them.