U.S. Productivity Surge Driven by Capital Use, Not AI Adoption

American businesses are achieving significantly higher output per worker, but the productivity boom appears driven by more intensive use of existing equipment and facilities rather than artificial intelligence deployment, according to new research from a major payments company.

Labor productivity has climbed to 2.5% annual growth over the past year, substantially above the 1.6% average of the previous two decades. While that difference may appear modest, sustained improvement at this pace would compound rapidly, meaningfully raising worker incomes and economic output within just a few years.

Why it matters

The distinction between capital utilization gains and technology-driven productivity growth has major implications for whether recent economic improvements will persist. Squeezing more output from existing factories and data centers has natural limits, while transformative technology could deliver compounding returns for years.

The capital utilization story

Ernie Tedeschi, chief economist at Stripe, examined what's actually driving the productivity numbers. His analysis reveals that total factor productivity—which measures output relative to both labor hours and capital deployed—has remained essentially flat even as labor productivity climbed.

The gains instead reflect higher utilization of assets already in place: factories running longer shifts, server infrastructure and GPU clusters operating at higher capacity, and hotel rooms achieving better occupancy rates. Economists categorize these improvements as capital intensity or utilization gains.

Tedeschi's cross-industry analysis found that sectors with high AI adoption rates do show stronger productivity growth, but this pattern existed before the pandemic and predates widespread use of large language models. The timing suggests factors other than recent AI advances are responsible.

Real gains with different implications

Higher capital utilization represents genuine economic improvement—companies are generating more value from their existing investments. However, Tedeschi distinguishes this from what he calls "microproductivity," the efficiency gains that emerge when new technologies fundamentally change how work gets done.

The analysis doesn't rule out future AI-driven productivity gains. Companies may still be working through implementation challenges, workflow redesigns, and organizational barriers that currently limit AI's economic impact. The technology's full productivity contribution could materialize as these obstacles are resolved.

Looking ahead

Tedeschi emphasized to the publication that the United States appears to be experiencing a genuine period of elevated productivity growth, with AI playing some role in the story. The critical question is whether AI represents a temporary factor or a more fundamental and lasting transformation.

Understanding the actual sources of productivity improvement will help economists and business leaders assess whether current growth rates can be sustained or whether they reflect a one-time adjustment as companies optimize existing capital stock.

These findings were first reported by Axios, based on Tedeschi's analysis of government productivity data and industry-level trends.