Analysis: US productivity rose by ~2.7% in 2025, nearly doubling the 1.4% annual average over a decade, as AI-exposed sectors cooled their entry-level hiring
New economic data suggests the US is transitioning to a phase of measurable gains from the technology
Financial TimesErik Brynjolfsson
Context & Ripple Effects
The reported productivity acceleration follows coverage that put the AI investment boom at about 1% of US GDP, making this an early indication that spending may be appearing in operating outcomes rather than only capital outlays. The earlier estimate of AI investment's GDP share provides the investment backdrop.
The labor side is already becoming more uneven: a later payroll analysis found shrinking employment for 22- to 25-year-olds in highly AI-exposed roles. That evidence of weaker young-worker employment makes entry-level hiring a key channel to watch alongside aggregate productivity.
First-order effects
AI-exposed employers are generating more output per worker while reducing the pace of entry-level hiring, shifting the immediate benefit-cost balance toward incumbent firms and workers with complementary skills.
The 2.7% productivity reading gives companies and investors a concrete benchmark for evaluating whether AI deployment is translating into measurable efficiency gains.
Second-order effects
Hiring plans may tilt toward experienced workers, implementation talent, and roles that complement AI, while traditional junior pipelines in exposed occupations face greater pressure.
The contrast with companies that spend heavily on AI and are adding workers faster than peers suggests that AI adoption need not mean uniform headcount cuts; outcomes may diverge by company type and ability to deploy the technology. AI-heavy companies' faster hiring sharpens that distinction.
Third-order effects
If productivity gains persist while junior hiring remains weak, firms may need to redesign training and early-career pathways rather than rely on the historical entry-level funnel.
The pattern points to AI industrialization becoming an operating-model issue, not just an infrastructure-investment cycle: measured output gains will increasingly be weighed against how broadly employment gains are distributed.
The trend: AI is moving from an investment-led story toward one in which productivity gains and the distribution of labor-market effects are tested together.
Millions of speculative articles on the effects of AI. Here is one making some data driven early conclusions, which are very positive in terms of the productivity boosting effects.
“For over a decade, economists have grappled with a modern iteration of the Solow Paradox: we have seen AI everywhere except in the productivity statistics ...Data released this week offers a striking corrective to the narrative” https://www.ft.com/... [image]
“Why am i paying for these kinds of slopvertorials to appear in this newspaper?” The comments in this hack piece are hilarious. FT readers are not having it.
Good article by @erikbryn on the gradual appearance of AI impacted productivity growth in the US economy. Let's hope the UK follows suit. https://www.ft.com/... via @ft
Boom. From my “living” post on AI and productivity last month, which noted that productivity gains in micro studies haven't shown up in the macro data yet. I guess “sooner” came pretty quickly. Updating it now (though as @EconBerger notes below, this is one data point). [image]
2.7% productivity growth in 2025, nearly double the previous decade. And this is likely companies deploying 2023-era models at scale. If the previous generation of AI is already moving the needle this much, what does 2026 look like when Opus 4.6 and GPT-6 hit real workflows?
Narrative violation! Giddyap. “General-purpose technologies, from the steam engine to the computer, do not deliver immediate gains. Instead, they require a period of massive, often unmeasured investment in intangible capital — reorganising business processes, retraining the [imag…
US productivity growth is likely to come in at about 2.7% for 2025. That is nearly double the average of the previous 10 years. There are many factors at work, but part of the story is that businesses are finally beginning to reap some of AI's benefits. I discuss the latest
Erik Brynjolfsson: For over a decade, economists have grappled with a modern iteration of the Solow Paradox: we have seen artificial intelligence everywhere except in the productivity statistics. Sceptics argue that the reason for this is that modern innovation in machine learni…
“We are transitioning from an era of AI experimentation to one of structural utility. We must now focus on understanding its precise mechanics. The productivity revival is not just an indicator of the power of AI. It is a wake-up call to focus on the coming economic transforma…